Structured Finance - Fitch Ratings · considers both the quantitative and qualitative factors listed in this criteria report when assigning ratings to auto loan ABS. ... Structured

www.fitchratings.com March 21, 2016

Structured FinanceAsset-Backed Securities / U.S.A.

Rating Criteria for U.S. Auto Loan ABS

Sector-Specific Criteria

Scope Analysis of Auto Loan ABS: This criteria report describes Fitch Ratings’ criteria and approach to analyzing new and existing auto loan-backed ABS transactions issued in the U.S. and Canada. Fitch considers both the quantitative and qualitative factors listed in this criteria report when assigning ratings to auto loan ABS.

Key Rating Drivers Collateral Performance: Fitch’s expectation of collateral performance is a key rating driver. Fitch assesses future performance expectations through an analysis of an originator’s historical static pool and securitization data, including delinquencies, defaults, net losses, recoveries, loss timing and prepayments. Fitch analyzes granular collateral pool/loan tape characteristics and risk concentrations to determine credit risks present that drive transaction loss frequency and severity.

Macroeconomic Risks: The economic environment can have a material impact on U.S. and Canadian auto loan ABS ratings. Fitch takes into consideration the strength of the economy, as well as future expectations, by assessing key macroeconomic indicators that are correlated with asset performance, such as unemployment rate.

Forward-Looking Approach to Derive Base Case Loss Proxy: Fitch derives its base case proxy by examining fully amortized historical vintage loss curves, extrapolating incomplete vintage curves, matching future expectations with any relevant previous vintage data, and weighting by relevant pool credit characteristics. Additionally, a margin of safety is built into the base case by weighting previous recessionary vintages and adding adjustments for future unemployment expectations and/or wholesale vehicle market conditions, where applicable.

Payment Structure: A transaction’s payment structure and cash flow allocations are major drivers in assessing credit enhancement (CE) adequacy and determining ratings. Fitch uses a Microsoft Excel-based internal cash flow model customized to reflect the transaction payment structure and tests the impact of stressing various assumptions, including prepayments, default timing, recovery rates and recovery lag. The output of the cash flow modeling is reviewed to assess whether the rated bonds are fully paid, in accordance with the transaction documents, in each stress scenario associated with a bond’s rating.

Legal Risks: Legal risks can affect the rating in the event that legal uncertainties pose a threat to the availability of cash flows or the collateral itself. Fitch’s legal analysis includes a review of the legal structure and legal opinions furnished by the originator to confirm cash flow derived from the assets will not be impaired (either as a result of the bankruptcy or insolvency of the originator or any other transaction party, such as a swap counterparty, or the trustee’s lack of a perfected first-priority security interest in the assets) or diminished (as a result of taxation).

Counterparty Exposures: Counterparty exposures can pose a risk to transactions if the relevant counterparties are a source of credit or performance weakness. Fitch’s analysis includes a review of counterparties in accordance with Fitch’s “Counterparty Criteria for Structured Finance and Covered Bonds” (Counterparty Criteria) (May 2014).

Seller/Servicer Operational Review: Disruption to servicing is a potential risk to the performance of auto loan ABS transactions. Fitch’s analysis includes a review of the corporate and operational activities (originations, underwriting and servicing) of the seller/servicer(s) to assess operational risks and the ability of the servicer to assume the servicing role for the transaction.

Inside This Report Page Scope 1 Key Rating Drivers 1 Data Adequacy and Compliance 2 Collateral Analysis 2 Credit Analysis 6 Structural Analysis and Cash Flow Modeling 16 Legal Analysis 24 Counterparty Analysis 25 Performance Analytics 29 Appendix 1 Additional Adjustments to Derived Loss Proxy 32 Appendix 2 Pool Collateral Data Requested 35 Appendix 3 Modeling Analysis Example (Hypothetical Trust XYZ) 41 Appendix 4 Cash Flow Modeling Sensitivity Analysis 43

This report replaces “Rating Criteria for U.S. Auto Loan ABS,” dated October 2015. This updated criteria report is substantially unchanged from the prior criteria, with minor changes/updates. Existing ratings will continue to be reviewed as part of the ongoing performance analytics process, consistent with the methodology of this report.

Analysts Hylton Heard +1 212 908-0214 [email protected]

Joyce Fargas +1 212 908-0824 [email protected]

Du Trieu +1 312 368-2091 [email protected]

John Bella, Jr. +1 212 908-0243 [email protected]

mailto:[email protected]




Rating Criteria for U.S. Auto Loan ABS 2 March 21, 2016

Structured Finance

Criteria Limitations

The criteria take into account the varying nature and degree of the collateral types and treatment thereof; the nuances surrounding regional markets and application therein, as well as legal frameworks; and the availability and applicability of relevant data. These may result in potentially varying degrees of analysis and applicability, and limitations thereof on a regional basis. In certain cases, such as data limitations, assumptions used in the rating process may be modified or supplemented as Fitch determines appropriate. Deviations from criteria will be disclosed in the applicable transaction rating report.

In terms of compliance with eligibility criteria, transaction documentation usually requires originators to repurchase any ineligible assets that could be intentionally or unintentionally sold to the issuer (covered in a transaction’s representations and warranties). Fitch assumes that originators will comply with such documented provisions and, therefore, Fitch’s credit analysis does not address the risk of ineligible assets being sold into the pool.

Data Adequacy The rating approach outlined in this criteria report uses historical performance data to form an expectation of future performance. Fitch expects to receive a minimum of five years of historical static gross (if gross is available) and net loss performance data, including a full economic cycle, to analyze a transaction. Static performance data include lender/issuer’s loss data from the total managed pool. Fitch expects historical data to be split into homogenous subgroups, where applicable.

Appendix 2 on page 35 lists data items Fitch expects to receive to analyze and rate a transaction. In the absence of certain data items, Fitch will assess the materiality and relevance of such data in accordance with its “Criteria for Rating Caps and Limitations in Global Structured Finance Transactions,” dated July 2015 (Rating Caps and Limitations Criteria). Depending on the outcome of such assessment, Fitch will either adjust its approach (e.g. not give credit to certain elements), cap the rating(s) or decline to rate the transaction. Where these alternative adjustments are applied, such approach and corresponding criteria will be highlighted in Fitch’s transaction rating report.

Collateral Analysis Fitch analyzes the characteristics of the collateral underlying a prospective auto loan securitization to determine the credit risk present in the specific pool of loans to be securitized. Fitch examines these collateral characteristics to better analyze the risks that drive both loss frequency and loss severity in a transaction. Fitch analyzes collateral characteristics utilizing collateral stratifications on a weighted average (WA) basis, and utilizes complete collateral loan tapes when provided by issuers and deemed applicable to utilize. Fitch will use these attributes to project losses for collateral characteristics that drive overall performance of an auto loan ABS securitized pool, as fully described in detail in the Credit Analysis section of this report starting on page 6.

Major collateral characteristics that Fitch analyzes and considers the most influential drivers of loss rates in auto loan ABS include the following: • Obligor credit quality (including internal and external credit scores). • Loan-to-value (LTV) ratios (when available). • Loan terms. • APRs.

Related Criteria Global Structured Finance Rating Criteria (July 2015) Counterparty Criteria for Structured Finance and Covered Bonds (May 2014) Criteria for Interest Rate Stresses in Structured Finance Transactions and Covered Bonds (December 2014) Criteria for Servicing Continuity Risk in Structured Finance (December 2015) Criteria for Rating Caps and Limitations in Global Structured Finance Transactions (May 2014)

https://www.fitchratings.com/creditdesk/reports/report_frame.cfm?rpt_id=867952



















Structured Finance

• Vehicle segment (car/truck/SUV) and condition (new/certified pre-owned/used). • Geographic diversification.

The vast majority of auto loans are fixed simple interest, level-pay installment loans payable monthly over a predetermined time frame, and with terms that typically range from 12−84 months. Each installment includes a portion of principal repayment plus interest and fees amortized over the life of the loan.

To conduct its analysis, Fitch expects to receive detailed loan/collateral loan tape and/or stratifications of the pool of loans being securitized. Appendix 2 (page 35) lists the loan stratifications/data Fitch expects to receive from an issuer.

Obligor Credit Quality and Scoring Fair Isaac Corp. (FICO) scores and internal proprietary obligor credit scores are generally two of the best predictors of defaults and loss rates of auto loans. The auto finance sector is typically segmented into three credit sectors: prime; midprime; and subprime (midprime and subprime are referred to as nonprime in this report). Fitch sorts the three sectors by an issuer’s historical cumulative net losses (CNLs); see the table below as a guideline to how Fitch classifies each sector.

Originators utilize proprietary systems that rank and assign scores to obligors based on expected default probability. Fitch utilizes FICO scores and/or internally derived scores during the credit analysis of a pool of loans, in combination with other credit-related stratifications. Many of these internal issuer models utilize internal credit scores combined with FICO scores, as inputs augmenting them with various predictive variables (for example: borrower’s current employment status; whether the borrower owns or rents a home and length of time therein; employment status/history; monthly income versus debt obligations; payment to income; and debt to income) in an attempt to determine the probability of obligor payment or default. Loan-to-Value Ratios Fitch examines LTV characteristics of the securitized loan pool, as well as an issuer’s managed portfolio and securitized pools, when this data is provided. Loans with high LTV ratios can result in increased loss severity due to lower equity at the time of default, and ultimately yield higher losses on defaulted loans. Fitch reviews static pool performance by LTV ratio when deriving its loss proxy for a pool, if available. Since most issuers do not provide loss performance stratified by LTV ratio, Fitch assesses the overall weighted average (WA) LTV ratio for the pool to determine consistency with the originator’s underwriting guidelines and

Auto Loan ABS Summary of General Collateral Characteristics Prime Midprime Subprime Credit Scores/FICO Commonly above 680 WA FICO Anywhere from 640−680 Mostly below 640 Loan LTV Above or below 100% More than 100% More than 100% Loan Term (Months) Greater than 60 is common;

pool may have a majority of longer term loans

Greater than 60 is common; pool may have a majority of longer term loans

Greater than 60 is common; pool may have a majority of longer term loans

Loan APR Below market Market/risk based Risk based Vehicle Mix (Majority) (New)/used (New)/used (Used)/new Geographic Mix Diverse Diverse/concentrated Diverse/concentrated Loss Experience Less than 3% CNL 3%−7% CNL Above 7% CNL

APR − Annual percentage rate. LTV − Loan-to-value ratio. CNL − Cumulative net loss.


Structured Finance

prior ABS transactions, and expects to receive LTV stratifications from an issuer (if available). Fitch uses other predictive collateral attributes, as described previously, in deriving a base case loss proxy for the pool, in the absence of LTV data.

LTV ratios on new vehicles are often defined as the loan balance as a percentage of the manufacturer’s suggested retail price as equipped or vehicle transaction cost, including dealer add-ons. For used vehicles, the value typically reflects the wholesale value determined by official price guides published periodically by Black Book USA, the National Automobile Dealers Association (NADA), or Kelley Blue Book. Underwriting guidelines typically include specified maximum advance rates for both new and used vehicles.

Loan Term The loan terms of the majority of securitized auto loans range from 12–84 months. Lenders may use extended-term loans, those with original terms longer than 60 months, to boost sales by lowering monthly payments for the obligors. Fitch evaluates the incremental risk associated with extended loan terms by reviewing and analyzing the performance of each loan term bucket of an issuer’s static portfolio and determining a loss proxy for each applicable loan term bucket, thereby deducing a WA loss proxy for the total pool of loans, when performance by term is available for each term bucket.

Currently, most issuers do not have or only provide limited loss performance data for certain extended loan term contracts, specifically loans with terms of 73–84 months. This is due to the fact that these types of loans are new to some issuers’ platforms and therefore may lack historical originations (five years of data and/or through a complete economic cycle). In cases where data may be limited for these extended loan term originations, Fitch can apply a loss timing curve approach during the derivation of the loss proxy, deriving timing curves based on loan terms (for example for terms up to 72 months) that have adequate origination history, and then extrapolate losses to derive loss projections on these longer-term loan concentrations (73+ month loan terms) in a pool.

Fitch will apply additive quantitative stresses to the portion of these extended loans present in a securitized pool, if the loss timing curve approach is deemed not applicable or in the absence of historical loss data, where deemed appropriate. Any deviations or new approaches utilized to assess credit risks present in such extended term loans will be discussed in Fitch’s rating report.


Structured Finance

Longer term loans have historically performed worse than shorter term loans. Longer term loans tend to have higher loss frequency and defaults, as these loans are more often underwritten to borrowers with weaker credit profiles. Additionally, there is a greater chance that the obligor’s credit could deteriorate during the life of the loan. Loss severity may also be greater for longer term loans, as they amortize slower than shorter term loans, and recovery proceeds on default may cover less of the outstanding principal balance.

The chart above highlights the affect longer term loans have on loss severity levels in the event of a default. A representative value depreciation curve for a midsize vehicle, combined with amortization schedules of 36-, 48-, 60- and 72-month contracts, is displayed above. The vehicle is assumed to sell for $21,000, the initial LTV is 110% and the interest rate is 7%. As illustrated, the negative equity position is considerably larger in the early months and lasts longer for a 72-month loan versus a 36-month loan.

Annual Percentage Rate Obligors are assessed APRs based on their credit profile and market interest rate trends. APRs are part of a lender’s risked-based pricing strategy obligors with weaker credit quality are charged higher interest rates, and vice versa. Fitch reviews the distribution of APRs within the loan pool. Although loans with lower APRs are generally extended to more creditworthy borrowers, they may decrease the availability of excess spread over time. Fitch reviews the credit quality of a pool and implicit pricing for each related credit tier. Loans with high interest rates are typically underwritten to nonprime obligors, and such loans have higher default probabilities than those underwritten to prime obligors.

Vehicle Type Different vehicle segments, brands and models experience different rates of depreciation. A pool of loans that exhibits make, model and year diversification produces more consistent depreciation rates, resulting in better recoveries on default and lower losses. Portfolios that lack make, model and year diversification are exposed to greater risks including changes in wholesale vehicle values, vehicle recalls or other quality-related problems and may cause a particular manufacturer’s vehicle or an entire segment to rapidly depreciate and negatively affect recoveries. If data is provided by vehicle type, Fitch examines static pool and securitization loss data to extrapolate losses as part of its loss proxy derivation for a transaction.


Structured Finance

Vehicle Condition (New/Used) Receivables in a securitized pool may be collateralized by both new and used vehicles. New vehicles usually comprise the majority of prime pools. Nonprime pools typically have a large majority of used vehicles due to weaker obligor credit profiles and the affordability factor of these vehicles. It is evident from static pool and securitization performance data that loans backed by used vehicles perform worse than new vehicle loans. Fitch reviews the vehicle mix of a collateral pool (new, pre-owned/manufacturer certified and used vehicles) and how it has changed over time, allowing the relative effect of compositional changes of an issuer’s portfolio to be measured. Fitch examines static pool and securitization loss data to quantify the performance differential between new and used vehicles, if available, and extrapolate losses as part of its loss proxy derivation for a transaction.

Geographic Diversification A pool of auto loan receivables that exhibits geographic diversification minimizes potential exposure to regional economic downturns. Most issuers have well-diversified portfolios with a national footprint. Fitch reviews the measures a lender employs to minimize portfolio exposure to local economic events. In cases of very high state concentrations, an adjustment will be applied to the base case loss proxy of a pool to mitigate this risk and further analyze the risk according to composition percentage and performance by each region. Any adjustments will be detailed in Fitch’s transaction rating report. An example of a geographic concentration stress is detailed on page 34.

Credit Analysis Introduction

Following the analysis and review of collateral characteristics in a pool of loans, Fitch derives expected base case cumulative gross loss (CGL) and/or CNL proxies for each collateral stratification (and/or subset thereof) to project the future performance of a transaction, as described in the Collateral Analysis section of this report. Upon derivation of these individually extrapolated loss proxies, each collateral subset is then weighted based on the composition of the pool to derive final CGL/CNL proxies for the entire securitized loan pool.

The loss proxy extrapolation is intended to extrapolate future transaction performance under similar conditions, including economic and auto industry conditions, utilizing previous portfolio and securitization performance of collateral with similar characteristics (if not similar, Fitch can apply additive stresses). Fitch then combines this loss proxy derivation by taking into account future economic and auto industry conditions by applying additive adjustments if deemed necessary. This is discussed in detail in the Loss Proxy Extrapolation Methodologies (listed on page 10) and Additional Adjustments to the Derived Loss Proxy (page 13) sections, encompassing forward looking criteria to project future performance.

Fitch then constructs a loss timing curve to evaluate the expected timing of losses utilizing historical static pool and/or securitization data. The timing of defaults and recoveries impacts cash flows available for an ABS transaction and, therefore, the ability of cash flows from the assets to repay the notes.

Fitch’s review of an issuer’s performance data and derivation of a loss proxy (loss derivation for the pool of loans) are primary components of the credit analysis of a securitization. The derivation of a base case loss proxy starts with the examination and extrapolation of an issuer’s historical static/managed loss data on a monthly, quarterly and/or annual vintage basis.


Structured Finance

Static pool data may be based on monthly, quarterly or annual originations, and frequently are stratified to analyze specific portfolio performance. Loss data may be stratified by various predictive collateral characteristics and commonly include stratifications by FICO and/or internal credit scoring; original loan term; new/used; subvented/nonsubvented (incentive versus no incentive) loans; and multivariate combinations of the various characteristics listed above. In instances where static pool loss performance data are not stratified into certain collateral characteristics or multivariate combinations are not available, Fitch will increase the weighting of certain other stratifications, utilize proxy data that is representative of future performance or otherwise make a qualitative adjustment to the derived base case loss proxy, if deemed appropriate, and will disclose these in a transaction rating report.

Fitch reviews the collateral and obligor characteristics of a pool and compares them with the issuer’s previously securitized pools and managed portfolio and static pool data, to gauge future performance (as discussed in the Collateral section). This aids Fitch in determining how loan characteristics may affect performance relative to prior pools of loans, and also looks forward to future performance assuming certain economic and auto industry conditions. Furthermore, by comparing securitized assets to an issuer’s total managed pool, Fitch is able to observe how the securitized pool should perform relative to the issuer’s overall managed portfolio in the future.

Static Pool Data Fitch utilizes static pool loss data as a primary data source in deriving base case loss assumptions and project future performance of a securitization pool. The terms vintage and static pool refer to a group of loans originated at approximately the same time on the basis of similar criteria/segmentation, the performance of which is tracked as a group over time (see the tables on page 8).

Use of Performance Data from Specific Historical Periods As detailed in the Key Rating Drivers on page 1, Fitch utilizes static portfolio data from certain periods or vintages to project future performance by capturing periods with similar economic and auto industry specific conditions that could occur during the life of a transaction. Fitch will evaluate performance data to determine the periods that are most reflective of future loss expectations, taking into account economic and auto industry (state of the wholesale vehicle market/used vehicle values) conditions predicted going forward for pools with similar collateral characteristics. If pools are not similar, then Fitch considers applying additive adjustments when deriving the loss proxy.

As mentioned previously, Fitch derives its base case proxy by examining fully amortized historical vintage loss curves, extrapolating incomplete vintage curves, matching future expectations with any relevant previous vintage data, and weighting by relevant pool credit characteristics. Additionally, a margin of safety is built into the base case by weighting previous recessionary vintages and adding adjustments for future unemployment expectations and/or wholesale market conditions, where applicable.

The periods deemed applicable and utilized to derive the proxy may be sequential, for example the 2006−2011 vintages, or may vary and be nonsequential period/vintage ranges (for example 2006−2009 and also 2011−2012), whichever is deemed most appropriate and best projects future performance of pools with similar characteristics under similar conditions. The specific periods/vintages utilized to derive the loss proxy will be detailed in Fitch’s transaction rating report.


Structured Finance

Securitized Pools of Loans In addition to the static pool analysis, Fitch reviews the historical performance of an originator’s historical securitized pools in order to project future performance of a securitization. Frequent issuers may provide static data derived from historical securitized pools for Fitch’s credit analysis. While the analytical process is similar to that of the static pool analysis, these results are interpreted with caution. Previously securitized pools may not be homogeneous and may be subject to changes in seasoning, overall asset characteristics and other factors, rendering the historical performance of those pools somewhat incomparable to each other and of limited use in deriving the loss proxy.

Sample of Quarterly Static Pool Data by Dollar Amount 3Q08 4Q08 1Q09 2Q09 Originations ($) 2,936,282,959 2,807,774,460 1,035,483,149 2,755,246,858 Loss Months ($) 1 9,279 2 118,513 70,350 46,699 138,702 3 463,985 496,282 490,159 523,814 4 1,465,319 2,012,424 2,004,447 1,615,214 5 2,928,825 4,110,684 4,043,635 2,306,605 6 5,700,632 7,838,530 7,649,016 7 8,813,451 11,485,469 10,748,808 8 12,649,694 15,572,318 12,214,592 9 16,977,819 20,400,169 10 21,035,485 23,521,141 11 25,345,461 24,619,610 12 28,914,266 13 30,905,008 14 31,952,832

Sample of Quarterly Static Pool Data by Percentage 3Q08 4Q08 1Q09 2Q09 Originations ($) 2,936,282,959 2,807,774,460 1,035,483,149 2,755,246,858 Loss Months (%) 1 0.0 0.0 0.0 0.0 2 0.0 0.0 0.0 0.0 3 0.0 0.0 0.0 0.0 4 0.0 0.1 0.2 0.1 5 0.1 0.1 0.4 0.1 6 0.2 0.3 0.7 7 0.3 0.4 1.0 8 0.4 0.6 1.2 9 0.6 0.7 10 0.7 0.8 11 0.9 0.9 12 1.0 13 1.1 14 1.1


Structured Finance

The analysis of the performance of securitized pools provides valuable information on losses, recovery rates and the overall performance of the issuer’s securitization program. It also provides a good indication of future securitizations’ performance assuming similar conditions.

Fitch has observed that some securitized auto loan pools perform better than the static portfolio with similar credit characteristics. The difference is primarily driven by a securitized pool’s eligibility criteria, which may require such limitations as a minimum number of payments collected from the obligor, eliminating early defaults. Some sponsors also elect to securitize assets in a narrower credit spectrum than the overall book of originations (i.e. a company may securitize only prime or new receivables, even though it originates nonprime and used receivables).

The analysis of historical loss performance of previously securitized pools is especially important when the securitized assets are significantly different from those of the managed portfolio.

If deemed applicable due to differences in static pool performance versus securitization performance, Fitch factors securitization performance projections into its analysis when deriving the base case loss proxy. An example of when securitization performance is utilized is if the lender changes origination and/or underwriting strategies or standards, and the performance of securitizations better reflects this change in terms of loss levels, while static pool performance differs or does not reflect such change as quickly.

When securitization performance is utilized to derive a loss proxy, Fitch will detail its approach in rating reports.

Base Case Loss Proxy Base case loss proxy assumptions are derived by Fitch utilizing quantitative and qualitative methods as discussed previously. Fitch derives five main assumptions utilizing historical static and/or securitization pool data to determine base case model inputs. These include the following: • CGLs. • Recovery rates. • CNLs. • Timing curves (default/loss). • Seasoning adjustment.

Fitch extrapolates less seasoned static pool vintages from loss speeds experienced by older vintages, as discussed in the Loss Proxy Extrapolation Methods section on page 11. The data may be volatile if there is existing volatility in seasoned vintages, and extrapolation may be skewed by less-seasoned vintages performing outside historical trends. However, Fitch takes into account these potential variations during its analysis as discussed in the remainder of this section, including the Additional Adjustments to the Derived Loss Proxy section.

Gross Losses Versus Net Losses Fitch will review static pool data on both a gross and net loss basis, when provided, although it often receives loss data only on a net basis. When gross and net loss data are available, Fitch can review historical loss rates and then determine recovery rates by calculating the difference between gross and net data. This enables Fitch to derive a base case recovery rate, which can be applied during loss extrapolations and cash flow modeling.


Structured Finance

If Fitch receives only historical data based on gross losses, it will extrapolate the data to determine a base case CGL projection and then derive a recovery rate assumption using historical recovery rates. Once both the CGL and base case recovery rates are projected, the pool’s CNL is calculated:

[CNL = CGL x (1 - recovery rate)]

The majority of the static pool data sets provided are composed of net data. Therefore, Fitch extrapolates CNL directly based on analysis of historical static pool data and derives the expected CGL by dividing historical base case CNL with the projected historical recovery rates [CGL = CNL/(1 - recovery rate)].

Regardless of the data set, the extrapolation methods described below are identical for both gross and net extrapolations. The specific data sets utilized in deriving Fitch’s base case proxy will be detailed in Fitch’s rating report.

Loss Proxy Extrapolation Methods Fitch extrapolates securitized and static pool default and loss data segmented for most major collateral characteristics detailed in the Collateral Analysis section, beginning on page 2, along with combinations of these various characteristics. Although the majority of static pool data are segmented by various pool characteristics, the analytical approach does not vary. Loss extrapolations are repeated for each subsegment to derive subpool-specific loss proxies for each one. The calculated proxies are weighted to mirror the collateral composition of the pool to establish a base case loss proxy for the pool.

Fitch utilizes two extrapolation methodologies, or a combination of both, to extrapolate a base case loss proxy. The methods are referred to as the nominal change method and percentage change method. While both methodologies look to the historical loss curves and speeds to project loss expectations of the more recent originations, they vary in analytical approach. The nominal change method produces more conservative results (higher loss extrapolations) when more recent vintages show improving performance, while the percentage change method is more reactive when projecting vintages with fewer data points and produces more conservative results when more recent vintages show deterioration.

Fitch compares the results of both projections to assess trend analysis of expected loss performance of the more recent vintages. Such analysis enables Fitch to identify more recent performance trends and determine any appropriate qualitative and/or quantitative adjustments to the projected loss extrapolation. The extrapolation method along with any adjustments incorporated in the loss extrapolations will be detailed in Fitch’s transaction rating report.

For the sake of simplicity, the descriptions of the two extrapolation methods will refer only to CNL, although the methods for CGL are the same as for CNL. In both methods, Fitch derives the historical monthly change for each data period for static pool vintages that have at least five data points using both the nominal change extrapolation method and the percentage change extrapolation method, as described on the next page.

Nominal Change Method • Under the nominal change method, CNLs are expressed as a percentage of originated principal

balance for each vintage. Each CNL data point is subtracted from that of the previous period. • The resulting data points of the same period across all vintages (e.g. month nine) are averaged to

calculate the nominal change factor (NCF).


Structured Finance

• The derived NCF for each period is then added to the corresponding CNL for each vintage in that period, that has not reached the end of the loss curve, until the vintage is projected fully to the expected loss curve.

Percentage Change Method • Under the percentage change method, CNLs are expressed in dollars. If the CNL data were

provided as a percentage, the percentage is multiplied by the origination volume for that vintage to calculate the CNL in dollars.

• The dollar amount of losses for all vintages with losses in a given period (e.g. month nine) are summed together and then divided by that of the prior period (e.g. month eight) to arrive at the percentage change factor (PCF).

• The derived PCF for each period is then multiplied by the corresponding CNL for each vintage that has not reached the end of the loss curve until the vintage is projected fully to the expected loss curve.

In both cases, all projected and actual CNL for each vintage are then averaged together to derive a historical average base case CNL assumption. The following example steps through the calculations for both methods to demonstrate how CNL curves are extrapolated. The methodology utilized to derive the loss proxy will be detailed in rating report.

The sample loss data above show the origination volume and CNL of four quarterly periods of originations of the subset of an auto loan lender’s overall originations. In this example, the extrapolation methodology will be illustrated using periods eight and nine from third-quarter 2008. For the nominal loss extrapolation method, the difference between period nine and eight is taken yielding an NCF of 0.15% (0.58% − 0.43% = 0.15%). For the percentage loss extrapolation, the PCF equals 133.70%, which is the sum of the latter period CNL divided by the sum of the prior period CNL (($17,030,441 + $20,496,754) / ($12,626,017 + $15,442,760) = 133.70%). Under both methods, the NCF and PCF are calculated for each period in each quarter of originations.

Historical Data-Driven Base Case CNL Derivation Methodologies (%)

3Q08 4Q08 1Q09 2Q09 Originations ($) 2,936,282,959 2,807,774,460 2,854,378,031 2,755,246,858 CNL Months 1 0.00 0.00 0.00 0.00 2 0.00 0.00 0.00 0.01 3 0.02 0.02 0.02 0.02 4 0.05 0.07 0.07 0.06 5 0.10 0.15 0.14 0.08 6 0.19 0.28 0.27

7 0.30 0.41 0.38 8 0.43 0.55 0.43 9 0.58 0.73

10 0.72 0.84 11 0.86 0.88 12 0.98

13 1.05 14 1.09


Structured Finance

The next step below takes the average of all the NCFs or PCFs for a given period to derive the historical loss curve.

The left half of the table above shows all NCFs for the sample data set, and the right half of the table shows all PCFs for the same data set. In both cases, the historical loss curve is calculated by taking the average of all change factors for any given period. As shown above left, the average NCF for period nine is 0.16%, and, as shown above right, the average PCF for the same period is 133.70%. Taken together, all the averages, or the derived loss curve, are then used to project the losses of the less seasoned vintages. This application is illustrated in the following table.

Applying the derived historical loss curves from the nominal method and the percentage method yields the resulting loss extrapolations depicted above left for the nominal method and above right for the percentage method. Using these extrapolation methods, the loss extrapolation for this sample segment of an auto loan pool has a loss proxy of 1.07% using the nominal method and 0.96% using the percentage method. Similar loss extrapolations would be generated for each subsegment of the pool, and then a WA loss proxy for the entire pool would be calculated.

Part One: Deriving Historical Loss Curves Nominal Method

Percentage Method

3Q08 4Q08 1Q09 2Q09 Average 3Q08 4Q08 1Q09 2Q09

% Change of Total Loss

for Period Originations ($) 2,936,282,959 2,807,774,460 2,854,378,031 2,755,246,858 Originations ($) 2,936,282,959 2,807,774,460 2,854,378,031 2,755,246,858 Monthly Variance Analysis (%) Monthly Variance Analysis ($) 1 0.00 0.00 0.00 0.00 0.00 1 0 0 0 0 2 0.00 0.00 0.00 0.01 0.00 2 0 0 0 275,525 0.00 3 0.02 0.02 0.02 0.01 0.02 3 587,257 561,555 570,876 551,049 824.15 4 0.03 0.05 0.05 0.04 0.04 4 1,468,141 1,965,442 1,998,065 1,653,148 312.00 5 0.05 0.07 0.07 0.03 0.06 5 2,936,283 4,211,662 3,996,129 2,204,197 188.41 6 0.09 0.13 0.13 0.12 6 5,578,938 7,861,768 7,706,821 189.76 7 0.11 0.13 0.11 0.12 7 8,808,849 11,511,875 10,846,637 147.38 8 0.13 0.15 0.05 0.11 8 12,626,017 15,442,760 12,273,826 129.44 9 0.15 0.17 0.16 9 17,030,441 20,496,754 133.70 10 0.14 0.11 0.13 10 21,141,237 23,585,305 119.18 11 0.15 0.04 0.10 11 25,252,033 24,708,415 111.70 12 0.12 0.12 12 28,775,573 113.95 13 0.07 0.07 13 30,830,971 107.14 14 0.04 0.04 14 0 0 0 0

Part Two: Applying Historical Loss Curves to Unfinished Vintages Nominal Method

Percentage Method

3Q08 4Q08 1Q09 2Q09 Average 3Q08 4Q08 1Q09 2Q09 Average

Originations ($) 2,936,282,959 2,807,774,460 2,854,378,031 2,755,246,858 Originations ($) 2,936,282,959 2,807,774,460 2,854,378,031 2,755,246,858 Vintage CNL Projections (%) Vintage CNL Projections (%) 1 0.00 0.00 0.00 0.00 0.00 1 0.00 0.00 0.00 0.00 0.00 2 0.00 0.00 0.00 0.01 0.00 2 0.00 0.00 0.00 0.01 0.00 3 0.02 0.02 0.02 0.02 0.02 3 0.02 0.02 0.02 0.02 0.02 4 0.05 0.07 0.07 0.06 0.06 4 0.05 0.07 0.07 0.06 0.06 5 0.10 0.15 0.14 0.08 0.12 5 0.10 0.15 0.14 0.08 0.12 6 0.19 0.28 0.27 0.20 0.23 6 0.19 0.28 0.27 0.15 0.22 7 0.30 0.41 0.38 0.31 0.35 7 0.30 0.41 0.38 0.22 0.33 8 0.43 0.55 0.43 0.42 0.46 8 0.43 0.55 0.43 0.29 0.42 9 0.58 0.73 0.59 0.59 0.62 9 0.58 0.73 0.57 0.39 0.57 10 0.72 0.84 0.71 0.71 0.75 10 0.72 0.84 0.68 0.46 0.68 11 0.86 0.88 0.81 0.80 0.84 11 0.86 0.88 0.76 0.52 0.76 12 0.98 1.00 0.93 0.92 0.96 12 0.98 1.00 0.87 0.59 0.86 13 1.05 1.07 0.99 0.99 1.03 13 1.05 1.07 0.93 0.63 0.92 14 1.09 1.11 1.04 1.03 1.07 14 1.09 1.12 0.97 0.65 0.96


Structured Finance

Additional Adjustments to Derived Loss Proxy In certain cases, Fitch may determine that the performance displayed in historical data are not representative of what can be expected in the current or near-term environment due to various factors, including changes in the business, industry or macroeconomic environment. In these cases, loss expectations will be adjusted to reflect trends and realities of the current and future environment in order to reflect evolving conditions that could impact asset performance. The adjustments listed below are applied during the derivation of the pool’s base case loss proxy in order to reflect future evolving factors that impact asset performance. These adjustments do not influence the loss multiple applied during cash flow modeling at any rating. Factors that may warrant adjustments include, but are not limited to: • Auto industry-related factors (e.g. loss severity adjustments such as wholesale

vehicle value adjustments). • Changes to domestic macroeconomic conditions (e.g. notable changes in unemployment). • Characteristics of the collateral pool or material changes to underwriting or origination

policies, including loan terms, LTVs and geographic diversification.

These adjustments are discussed in more detail in Appendix 1, and will all be discussed in a transaction rating report, if applicable.

Default Timing The timing of defaults through the life of a transaction has a significant impact on cash flows; therefore, it is an important assumption of cash flow modeling. Additionally, default timing curves are used to determine the seasoning adjustment (if any) to the base case loss proxy. Fitch utilizes those timing curves, adjusted if necessary, when modeling the expected cash flows (see the Structural Analysis and Cash Flow Modeling section beginning on page 16 for more details).

Default distributions differ by a consumer’s credit profile, LTV, loan terms, vehicle age and other pool characteristics. Therefore, the expected default timing curve is unique for each issuer. Fitch formulates its default distribution curves utilizing historical data of static pools or prior securitizations, or a combination of both, when applicable. Similar to deriving the subpool-specific, weighted, base case loss proxy, loss timing curve projections of each subpool may be aggregated according to the pool weight for each subpool (see the Loss Distribution Curve Derivation Example table on the next page).

Sample Base Case Derivation An Example Vehicle Condition Credit Score Term (Months) Pool Composition (%) CNL Proxy (%) New 700+ 61−72 50 1.75 New 700+ 0−60 20 0.85 New 620−700 61−72 10 3.50 New 620−700 0−60 5 1.50 Used 700+ 0−60 10 1.20 Used 620−700 0−60 5 2.50 Expected Unadjusted WA Base Case CNL 1.72


Structured Finance

The loss expectation of the seasoned pool is different from that of the unseasoned pool because a portion of losses would have been realized on the seasoned collateral, while the unseasoned pool would not have experienced those losses yet. The default timing curve changes correspondingly, and Fitch will adjust the derived default timing curves for highly seasoned collateral pools, if such credit is warranted. The various timing curves will be detailed in a transaction rating report.

Occasionally, Fitch has observed that the default curves of the previously securitized trusts are significantly different than those of the static pool data. The variance can be explained by the collateral eligibility criteria and the structural features of the securitized trusts. If there is a significant variance in the default curve timings of the static pools and securitized trusts, Fitch will make an adjustment to the derived default curve, where deemed appropriate.

Fitch determines a base case loss timing curve for each transaction based off historical portfolio and/or securitization loss curves. In addition to the base case loss timing curve, Fitch also conducts sensitivity analysis around two differing loss timing curves. In total three curves (front, mid and/or back-ended) for each transaction are evaluated to see the impact on loss coverage levels and rating multiples. The results of the sensitivity analysis will be considered by the credit committee. If any of these additional curves or adjustments are utilized in the rating analysis, they will be detailed in the applicable transaction rating report.

Seasoning Adjustment The derived base case loss assumption does not account for seasoning of the collateral. The expectation of the losses for an outstanding balance may differ from the projected base case loss assumption based on the timing of defaults and the vintage’s amortization speed. To compensate for this, Fitch will make an adjustment to the calculated base case loss based on the remaining losses as a percentage of the outstanding pool. The adjustment will not be made

Loss Distribution Curve Derivation Example

New Used Pool

Derived Loss

Curve (%)

Loss Per Annum

(%) Term (Months) 48 60 72 48 60 72 62.76 Base Case CNL (%) 2 3 5 4 5 8 4.09 Pool ($ Mil.) 100 300 400 100 70 30 1,000

Months 1 0.00 2 0.10 0.15 0.20 0.10 0.12 0.67 1.64 3 0.10 0.25 0.45 0.35 0.27 0.20 1.62 3.96 4 0.25 0.40 0.90 0.50 0.34 0.25 2.64 6.45 5 0.35 1.01 1.60 0.90 0.50 0.28 4.64 11.34 12 0.80 2.70 5.00 1.80 1.05 0.60 11.95 29.22 29.22 24 1.50 5.85 10.00 3.40 2.28 1.20 24.23 59.24 30.02 36 1.80 7.65 13.00 3.80 2.98 1.56 30.79 75.28 16.04 47 1.98 8.51 15.70 3.99 3.28 1.90 35.36 86.45 48 2.00 8.55 16.00 4.00 3.33 1.92 35.80 87.53 12.25 49 2.00 8.60 16.40 4.00 3.35 1.95 36.30 88.75 59 2.00 8.99 18.70 4.00 3.49 2.25 39.43 96.41 60 2.00 9.00 19.00 4.00 3.50 2.28 39.78 97.26 9.73 61 2.00 9.00 19.20 4.00 3.50 2.30 40.00 97.80 71 2.00 9.00 20.00 4.00 3.50 2.39 40.89 99.98 72 2.00 9.00 20.00 4.00 3.50 2.40 40.90 100.00 2.74

Note: Resulting loss speed curve: 29%/30%/16%/12%/13%.


Structured Finance

when transactions have a revolving period, unless the eligibility criteria include a minimum level of seasoning. The formula used to make the adjustment is outlined in the example starting in the paragraph below.

As an illustration, assume the portfolio has six months’ seasoning and Fitch’s expected base case loss is 1.72%. Based on the historical loss timing curve, the loss percentage at month six is approximately 0.20%, while the historical pool factor at six months equals 95% of the original principal balance. In such a scenario, the base case loss would be adjusted downward by 0.12% to 1.60%, as calculated below.

(1.72% – 0.20%) / 95% = 1.60%

This downward seasoning credit adjustment reflects the fact that the portfolio amortizes more slowly than the defaults accumulate, compared with the cumulative base case. Of the losses, 11.63% was incurred, while 5.00% of the pool was amortized. If, on the other hand, 20.00% of the pool was amortized resulting in a pool factor at 6 months of 80%, the appropriate seasoning adjustment would increase the base case loss by 0.18% to 1.90%.

(1.72% – 0.20%) / 80% = 1.90%

In instances where origination vintage pool factor data are not available, Fitch utilizes a conservative prepayment assumption, keeping in mind historical performance to solve for a vintage pool factor, including defaults and voluntary prepayments. Seasoning adjustments will be noted in the transaction rating report.

Recovery Rates Recoveries from the sale of repossessed vehicles are a key determinant of loss severity in auto loan ABS. Additional collections may come from other sources, such as insurance claim payments for total loss salvage, nonrepairable or stolen vehicles, deficiency balance collection efforts or sales and/or credit default insurance. Since most repossessed vehicles are sent to used vehicle auctions for disposition, proceeds from auction sales are the largest component of recoveries. These amounts are applied (by issuers) to reduce the charged-off balance, with all repossession and disposition costs netted out of realized proceeds. Recovery rates are influenced by several interrelated factors, each of which has specific drivers: • The macroeconomic environment. • The new vehicle market. • Wholesale and auction market conditions. • Vehicle type, condition, age and location. • Underwriting.

Seasoning Adjustment An Example

Seasoning Adjustment Formula Seasoning Adjusted Loss Proxy Equals: (Unadjusted Loss Proxy-Derived Cumulative Loss Rate at Pool’s Seasoning) (Historical Pool Factor at Pool’s Seasoning)

Example Two 20% Pool Amortization Portfolio Seasoning (Months) 6 Fitch Base Case Loss Proxy (%) 1.72 Loss % at Month Six 0.20 Historical Pool Factor at Month Six (%) 80.00 Base Case CNL Adjustment (%) 0.18

Example One 5% Pool Amortization Portfolio Seasoning (Months) 6 Fitch Base Case Loss Proxy (%) 1.72 Loss % at Month Six 0.20 Historical Pool Factor at Month Six (%) 95.00 Base Case CNL Adjustment (%) (0.12)


Structured Finance

• Servicing.

An originator’s recovery rate data are analyzed through aggregate portfolio statistics, performance of prior securitized pools and historical static pool data. Results will naturally vary, depending on the loan characteristics, including term, LTV, vehicle age, type and condition. Recoveries are influenced by external factors, such as repossession timing, disposition costs, wholesale auction values and economic conditions. If recovery data are limited, Fitch compares an issuer’s historical recovery rates with historical and current wholesale data segregated by vehicle segment/type (car/SUV/truck), brand and model to derive an expected recovery assumption. If no recovery rate data are provided, Fitch utilizes available market sources for recovery rate data. The analysis of all these variables leads to the historical base case recovery projection for a securitization.

Structural Analysis and Cash Flow Modeling Fitch analyzes a series of stress scenarios to determine the rating of a given tranche of notes and whether the payment of interest and principal is made fully according to the terms and conditions of transaction documents across all scenarios and can withstand cash flow stresses and payment shortfalls commensurate with the various rating categories. While the cash flow model addresses complex transaction features and is an important consideration in determining the final rating, ratings are ultimately assigned by a Fitch credit committee, which takes into consideration both quantitative and qualitative factors. The analysis entails: • Modeling the asset side/collateral characteristics. • Modeling the transaction’s liability structure. • Conducting sensitivity analyses.

Regardless of the structure utilized, Fitch customizes its internal cash flow model, the ABS Amortizing Loan Model, to replicate the flow of funds outlined in the transaction documents. This customization includes the incorporation of liability-side payment priority, yield supplementing mechanics and discounting of collateral and structural triggers, where applicable. Fitch’s ABS Amortizing Loan Model provides a uniform platform for the analysis of a wide range of transactions to compare the results in a consistent manner.

The main variables feeding the cash flow model are the portfolio scheduled amortization profile, default and recovery assumptions, prepayment assumptions, portfolio yield, capital structure and interest rates. Scenarios modeled by Fitch may include the following assumptions: • Primary historical default timing and two additional curves as previously described (Front-,

mid- and back-loaded default timing). • Historical prepayment rates. • Increasing, stable and decreasing interest rates (only for transactions with floating-

rate coupons).

The scenario ultimately utilized in the analysis will be determined by the credit committee and will be detailed in the applicable transaction rating report.

For certain transactions, other scenarios may be tested to identify the sensitivity of the transaction to different assumptions. Fitch’s cash flow model reflects how various stress scenarios affect principal and interest proceeds as they are received through the life of a transaction. The cash flow model then allocates those payments to the various classes of notes, based on the transaction structure as detailed in the underlying documents. If the cash flow model shows that a particular class of notes has received principal and interest payments, according to the terms and conditions of the notes under the stress scenario for a particular rating, then it is deemed to have passed that particular stress scenario.


Structured Finance

Representative Lines of Collateral The process of evaluating and incorporating asset characteristics and risks into realistic cash flow scenarios poses a more significant challenge than simply assigning the liability cash flows according to the waterfall. To successfully model payment streams from a pool of receivables, Fitch utilizes individual representative pools (replines) of the aggregate pool based on collateral and performance-based characteristics. Subsets of a pool with differences in the key characteristics, such as those listed below, may warrant segmenting collateral into replines: • APR. • Original term. • Remaining term. • Prepayment rate. • Loss severity. • Loss assumption. • Loss timing curve. • Vehicle condition.

The following example illustrates how creating separate replines of a collateral pool is important to properly assessing the risks of an aggregate pool. The hypothetical pool is from a captive finance company and contains equal concentrations of 2% APR subvented collateral and 10% APR market-rate collateral. The original term and remaining term of the two underlying pools are equal. Thus, absent prepayments on the collateral, the difference in cash flow between a single pool with a 6% APR and two pools equally sized with 2% and 10% APRs would be minimal.

However, the actual prepayment experience between subvented (incentivized loan rates) and market-rate collateral is substantial, both from a voluntary standpoint, as subvented rate obligors have little incentive to pay off a below-market-rate loan, and from an involuntary

perspective, since better credit-quality obligors who qualify for incentive loans have less propensity to default on their loans. As a result, the historical experience for the hypothetical issuer points to a 0.50% absolute prepayment speed for the subvented collateral and a 1.80% absolute prepayment speed for the market-rate collateral, compared with the overall prepayment speed of 1.20%.

Representative Lines of Collateral (Replines) An Example Total Pool Subvented Rate Pool Market Rate Pool Balance ($) 10,000,000 5,000,000 5,000,000 Original Term (Months) 60 60 60 Remaining Term (Months) 55 55 55 APR (%) 6 2 10 Prepay Assumption ABS (%) 1.20 0.50 1.80

ABS − Absolute prepayment speed.


Structured Finance

The chart below compares the two scenarios discussed above; the remaining principal balance of the loans is represented by the bars, and the outstanding APR of the total pool is represented by the lines. The principal amortization of the collateral under both scenarios is very similar; however, the total APR of the pool is significantly different in the two repline examples. As shown in the chart, lower interest payments will be received for the single pool.

Fitch reviews static pool data and securitization prepayment history. Prepayments will be applied in the cash flow model based on collateral composition and historical performance.

Liability Structure Fitch reviews the liability structure of transactions that are proposed by originators and utilizes its internal amortizing loan cash flow model to evaluate the proposed structure. It should be noted that Fitch does not recommend or approve any particular structural features.

Revolving Structures U.S. and Canadian auto loan ABS transactions may feature a revolving period, during which, for a specified period, principal collections are used to purchase additional receivables rather than to repay notes. Revolving periods expose noteholders to additional risks with respect to a longer risk horizon and the pool asset quality and performance, than traditional fixed-collateral pool auto ABS transactions. Fitch evaluates and addresses these risks present in transactions through its analytical assumptions and stresses. However, longer revolving periods could expose noteholders to the risks of a substantial change in the originator‘s underwriting criteria and the resulting portfolio deterioration. Fitch may decline to rate or may cap the maximum achievable rating of transactions that have long revolving periods.

The risk of performance deterioration during a revolving period may be partially mitigated by the presence of eligibility criteria, amortization triggers, including adequately sized performance-based triggers, CE-based triggers, asset-level triggers and other triggers. In revolving transactions, large amounts of cash may be held pending the purchase of additional receivables. Fitch expects counterparty risk to be sufficiently mitigated, either structurally or by ensuring counterparty risks are addressed adequately, thus meeting Fitch’s Counterparty Criteria. Furthermore, funds held in accounts are expected to be invested in permissible investments also outlined in the Counterparty Criteria.

Fitch will monitor revolving transactions by reviewing transaction performance data, compliance with performance triggers, and changes to pool compositions throughout the revolving period.


Structured Finance

Fitch expects to receive from specific collateral and/or performance data from issuers dictated by eligibility criteria and concentration limitations throughout the life of the transaction, as deemed necessary to monitor asset performance. Revolving transactions are exposed to the risk of changes in origination standards during the revolving period. In the event that collateral and/or performance deterioration is reported or foreseeable, Fitch will request ongoing updates to items listed above as well as originations, underwriting and servicing updates or changes in strategies, policies or procedures on an operational level. Additionally, the agency expects originators and servicers to provide prompt notification of any material changes to their origination, underwriting or servicing processes on a go forward basis.

Fitch will derive a base case CNL proxy determined for the initial collateral pool, consistent with the methodology discussed in the Base Case Loss Proxy section starting on page 9. Then Fitch will utilize stated maximum collateral concentration and collateral concentration limits to form a WA base case CNL, after the revolving period, thus assuming the pool migrates to the worst case possible pool under the concentration limits during the revolving period.

Credit Enhancement CE is typically provided in the form of subordination, overcollateralization, cash reserves and excess spread, including synthetic yield structures like yield supplement accounts (YSAs) and yield supplement overcollateralization accounts (YSOCs). Fitch’s “Global Structured Finance Rating Criteria,” dated July 2015, available on its website at www.fitchratings.com, discusses the various forms of CE and risks inherent in each. Fitch will review the relevant CE structure of each transaction and apply it within the agency’s cash flow model.

Excess Spread Considerations Where the portfolio yield exceeds interest costs and expenses, excess spread will be a source of funds to cover defaults, subject to the transaction priority of payments. Excess spread is considered by Fitch during cash flow modeling, according to the transaction structure/waterfall. The availability of excess spread to cover defaults will depend on the prepayment, yield and delinquency performance of receivables.

Many auto loan transactions rely heavily on excess spread to build other forms of CE, such as overcollateralization and cash reserves, to pre-specified target levels. Structures with a disproportionate reliance on excess spread may place some notes in a precarious situation, as CE may not build as expected due to higher than anticipated delinquencies and losses, increasing such structure’s susceptibility to adverse asset performance versus a structure with hard CE. Structures with heavy reliance on excess spread could be exposed to higher rating volatility relative to structures with similar target CE levels but composed of a greater proportion of hard CE.

Hence, Fitch places heavy emphasis on rating sensitivity analysis when evaluating structures with high reliance on excess spread. Fitch will adjust rating multiples and/or cap or assign lower ratings should classes of notes prove to be prone to greater rating volatility due to low levels of initial hard CE in the structure with high reliance on excess spread to cover potential losses. Any deviations from the traditional loss multiple coverage approach will be disclosed in the transaction rating report.

Fitch evaluates each class of notes’ reliance on excess spread for loss protection, particularly since excess spread may be less reliant for enhancement, depending on structural transaction mechanisms. This is particularly the case for junior, lower rated bonds (for example, BBBsf


Structured Finance

rated bonds). On closing, reliance on excess spread as enhancement is mostly at higher levels as a percentage of total enhancement. However, structural mechanisms present, including quick pool amortization and growing enhancement levels, reduce the reliance on excess spread to cover potential losses.

Furthermore, Fitch’s cash flow modeling assumptions, including prepayment assumptions, delinquent interest stresses and various loss timing speed assumptions, reduce the implicit credit given to excess spread. Fitch also considers a back-ended loss timing curve/scenario during its analysis of initial excess spread levels, since this scenario (unless spread trapping mechanisms exist) will result in lower loss coverage levels. This is due to excess spread releasing early in the life of a transaction.

Lastly, Fitch also considers certain qualitative factors where the reliance on excess spread is high overall, particularly for subordinate or junior bonds. This includes taking into account the strength of the originator/servicer, the nature of the collateral (such as prime or nonprime credit quality) and the historical performance and volatility of an issuer’s managed portfolio and securitizations.

Delinquent Interest Payment delinquencies result in a delay in cash flows available to pay noteholders, create liquidity concerns and decrease periodic available excess spread. In addition, late payments are an indicator of potential future defaults. Fitch reviews historical delinquency trends by reviewing prior securitization and historical managed portfolio delinquency performance. Fitch incorporates a delinquent interest stress into cash flow modeling, effectively reducing the amount of excess spread flowing through the cash flow model. The extent of delinquent interest stresses applied will be noted in the transaction rating report.

Servicing Fees Fitch incorporates the cost to service collateral in the cash flow model. In some cases, when the originator acts as the servicer, actual servicing costs in the transaction documents may not reflect the market rate a substitute servicer would expect to receive. In these circumstances, Fitch utilizes a more conservative market-rate servicing fee in the cash flow modeling scenarios.

Prepayments To derive prepayment assumptions for a transaction during cash flow modeling, Fitch reviews an originator’s historical prepayment rates, absolute prepayment speeds in auto loan ABS (commonly referred to as ABS speeds), to determine a base case prepayment speed. If deemed applicable, Fitch will stress this base case prepayment speed further to stress levels of excess spread. This is especially the case where historical prepayments exhibit volatility or where certain collateral characteristics are present in a pool (for example, a pool with a low WA APR that may not reflect historical prepayment speeds).

Fitch’s base ABS prepayment speed is applied in its cash flow model to reflect historical prepayment levels of an issuer, and can be further applied on an individual repline basis to utilize a more granular approach to reflect collateral-specific historical prepay trends. When data are available, Fitch reviews managed portfolio prepay speeds on an APR basis to determine whether to apply bifurcated prepayment speeds to specific portions of the pool, driven by APRs. Fitch’s prepayment assumptions are detailed in the transaction rating report.


Structured Finance

In certain cases where collateral pools have significant compositions of subvened (incentive APRs – very low APRs of 0% APRs) collateral, meaning low APR subvened contract rates or notable bar-belled APR compositions (for example, a high percentage of 0% APRs and high percent of very high APR contracts), Fitch will conduct an additional prepayment speed sensitivity analysis that will be considered by the credit committee. If any of these additional curves are utilized in the rating analysis, they will be detailed in the applicable transaction rating report.

Determination of Legal Maturity If the most senior class of notes is to be money market eligible, the legal final maturity dates for such class will be determined in accordance with Rule 2a-7 under the Investment Company Act of 1940 as the last payment date prior to day 397 following the pricing of the transaction. For all classes of notes that will not be money market eligible, and the most subordinate class in the capital structure, legal final dates are expected to be established by determining the date on which each specific class of notes will be paid in full when amortizing the pool with 0% prepayments and 0% defaults, then adding another three months to the result. For the most subordinate class of notes, the legal final date is expected to be established by adding six months to the longest contract maturity date to account for extensions and recoveries.

Money Market Tranche Sizing Fitch’s criteria for money market sizing are to amortize the pool with 0% defaults and a 0.0% absolute prepayment speed. The amount of collateral repaid in the first 13 months is the maximum allowable money market size to achieve a short-term rating of ‘F1+’.

Interest Rate and Currency Risk Fitch will identify any underlying interest rate or currency mismatches and analyze the extent to which these positions are mitigated through the transaction’s hedging structure, if any. Where interest rate or exchange rate (FX) risk is not fully hedged, Fitch may apply stresses in its cash flow model. Fitch expects any relevant hedge counterparties to be consistent with its hedge counterparty criteria detailed in Fitch’s Counterparty Criteria. If a transaction includes material exposure to unhedged interest rate or FX risk, Fitch may decline to rate the transaction.

Priority of Payments Fitch will review the priority of payments to identify the relative seniority of each obligation. Fitch will replicate the transaction-specific priority of payments within its cash flow model.

Note Amortization Different classes of notes will may be amortized either on a sequential or pro-rata basis. Under sequential amortization, principal funds are allocated first to repay senior notes in full, before being allocated to repay junior notes. Under pro-rata amortization, funds are shared in proportion to the note balance. Pro-rata structures are vulnerable to back-loaded loss profiles and tail risk. Structures may feature both sequential and pro-rata amortization phases, i.e. as a result of an event of default or breach of a trigger. The amortization sequence will be included in Fitch’s cash flow modeling.


Structured Finance

Auto Loan ABS Rating Stresses/Loss Multiples Fitch’s rating analysis includes stress multiples to incorporate the risk that the actual default performance may be worse than the base case loss assumption for the pool. This approach is intended to size the default risk that could occur under each rating stress. The stressed assumptions are intended to provide a basis to account for the impact of macroeconomic and industry-related deterioration on the transaction’s cash flows. The multiples are applied to the base case loss expectation to arrive at the stressed loss for a given rating scenario.

Fitch’s range of multiples for prime and nonprime auto ABS transactions is shown in the table below. In the prime sector, rating multiples applied by Fitch are 5.0x for ‘AAAsf’ rated notes, 4.0x for ‘AAsf’, 3.0x for ‘Asf’, 2.0x for ‘BBBsf’, 1.5x for ‘BBsf’, and 1.0x or higher for ‘Bsf’. The multiple ranges at each rating category incorporate the spectrum of possible ratings, including both the “minus” and “plus” rating modifiers. Nonprime transaction multiples applied by Fitch are lower than prime multiples given the nature of the asset class, structural features of transactions, significantly higher base case loss levels and lower volatility of delinquencies and losses.

Prime and nonprime multiples are differentiated on a practical basis. For example, a subprime transaction with a CNL loss proxy of 15% will have gross defaults of approximately 23%, assuming low recoveries of 35%. If a stress commensurate with ‘AAAsf’ ratings is applied utilizing the 4.5x low end of the typical prime scale, the target defaults would be an impossible 104% of the collateral pool, a level that is not practically possible, so it is limited to 100% of the pool. Loss multiples are also at least partially based on historical volatility. The high level of expected delinquencies and losses in subprime pools relative to prime collateral has contributed to significantly lower volatility in performance. Historically, subprime late-stage delinquencies have been 25%−30% less volatile, while subprime annualized net losses have been more than 50% less volatile than prime collateral.

Furthermore, Fitch understands that multiple levels may compress over time due to the underperformance of a transaction versus initial expectations and takes this into account on an ongoing basis. The multiple will be determined on a transaction-specific basis, in conjunction with the base case assumption, and may vary due to various qualitative factors, including: • Seller/Servicer Risks: If a transaction servicer is of weaker financial standing, has limited

servicing history or exhibits any other characteristics that could negatively influence the performance of the portfolio, a higher multiple may be employed.

• Quantity and Volatility of Historical Data: Limited data history and volatile performance may give rise to concerns regarding the consistency of the origination and collection processes. Such portfolios could likely be more volatile under economic stress; therefore, a higher multiple may be applied. However, notably, a higher stress cannot mitigate the provision of insufficient data.

• Stability of Origination Volumes: Historical data derived from periods of unstable, especially growing, origination volumes may be less predictive than data derived from a period of stable origination volumes. This is because the same factors that support increased origination volumes

Fitch's Rating Loss Multiples (x) Rating Category

Prime Rating Multiplesa

Nonprime Rating Multiplesa

AAAsf 4.50−6.00 3.00−4.50 AAsf 3.75−5.00 2.50−3.75 Asf 2.50−3.75 2.00−2.75 BBBsf 1.75−2.50 1.50−2.25 BBsf 1.25−1.75 1.25−1.75 Bsf 1.00−1.25 1.00−1.25 aThese represent typical Fitch rating multiple ranges for the prime and subprime sectors.


Structured Finance

may also negatively impact future performance (e.g. broader distribution and lower underwriting standards). This may result in a higher multiple being applied.

• Collateral Characteristics: Sufficient loss data may not be available for loan products that have not been historically offered to borrowers. The inclusion of these loans in securitizations may result in the use of a higher multiple.

Many transactions feature both positive and negative aspects from the above. The importance of these factors will vary for different portfolios. The final stresses applied will result from the qualitative and quantitative factors described above. The rationale for the stresses applied will be described in the transaction rating report.

While the above range of stresses provides a tool to reflect the expectation that different portfolios will respond differently to economic deterioration, notably, the application of higher stresses does not negate the importance of adequate origination and servicing practices, as well as the availability of sufficient historical data when setting base case expectations. Without such data, Fitch may be unable to derive base case expectations with a sufficient degree of comfort to apply this rating approach, and, in such event, a rating cap may be applied to the transaction (and will be noted in the transaction rating report). See Appendix 3 starting on page 41 for a detailed example of the typical cash flow modeling analysis Fitch implements and follows for an auto ABS transaction.

Rating Sensitivity Analysis In addition to the sensitivity analysis of cash flow modeling assumptions, Fitch analyzes the rating sensitivity of the notes to collateral performance outside of initial expected levels for each auto loan ABS transaction. As Fitch’s base case loss projections are derived primarily from historical collateral performance, cyclical increases in the frequency of defaults and loss severity on defaulted receivables could produce CNL levels higher than the base case. This will result in CE declines and remaining loss coverage levels available to the notes and may make certain note ratings susceptible to potential negative rating actions, depending on the extent of the decline in coverage.

Fitch’s rating sensitivity analysis consists of stressing a transaction’s initial base case loss assumption by 1.5x and 2.5x and examining the rating implications on all classes of issued notes. Due to the higher loss assumption in a subprime transaction, rating sensitivity is run at 1.5x and 2.0x for such transactions. Rating multiples applied during this sensitivity analysis are consistent with the credit analysis of the transaction. The 1.5x and 2.5x (or 2.0x for subprime pools) increases of the base case loss represent moderate and severe stresses, respectively, and should provide a good indication of the rating sensitivity of notes to unexpected deterioration of a trust’s performance.

Due to rapid de-levering and structural features, a typical auto ABS transaction tends to build CE over time. The greatest risk of losses to an auto ABS transaction is over the first one to two years of the transaction, where the benefit of de-levering may be muted. Therefore, Fitch will focus on coverage levels over the first 18 months of the transaction’s life to examine a structure’s ability to withstand the 150% and 250% stress loss scenarios.

During the sensitivity analysis, Fitch models cash flows with the revised loss proxies while holding constant all other modeling assumptions, such as prepayment speeds, chargeoff and recovery lags, recovery rates and loss timing curves. The resulting cash flows are examined to determine periodic CE in six-month intervals. The ensuing loss coverage levels and multipliers at each interval are analyzed, as their results determine the rating sensitivity of all rated notes.


Structured Finance

Fitch will run additional sensitivity scenarios if deemed applicable, including testing sensitivities on default timing.

For illustrative proposes, to the right is the sensitivity analysis of a hypothetical prime transaction with five classes of notes with initial ratings of ‘AAAsf’, ‘AAsf’, ‘Asf’, ‘BBBsf’ and ‘BBsf’. All base case, 1.5x and 2.5x stressed loss assumptions, along with commensurate rating levels, are detailed.

For each hypothetical class depicted in the rating sensitivity table, the 1.5x base case loss scenario, the loss multiplier compresses from the initial level by month six; however, the available expected excess spread and rapid de-levering of the structure slowly builds CE, increasing the multiplier over time into month 12, further increasing by month 18. In this example, the 1.5x loss stress is not likely to cause a rating downgrade of any class of notes, as the multiples increase to within the range of the respective initial rating for each class.

The 2.5x base case loss scenario further compresses the multiples beyond those of the 1.5x scenario. In this case, the de-levering of the structure does not build enough CE to offset the impact of the increased losses, and each class of notes in this example would likely see a rating downgrade of up to two rating categories.

Legal Analysis Fitch’s analysis of special-purpose vehicles (SPVs) in structured finance is described in detail in Fitch’s Global Structured Finance Rating Criteria. Retail auto loan securitizations are structured by the sponsor to isolate the auto loans from the bankruptcy or insolvency risks of the entities involved in the transaction, namely the auto loan originator and/or seller and the parent of the issuer. Refer to the aforementioned criteria report for further information on Fitch’s expectations for SPVs in structured finance transactions.

Rating Sensitivity Example (Ratio of Available CE to Projected Losses [Loss Coverage Multiples])

Six Months 12 Months 18 Months

Class A Notes (x) Base Case Loss Proxy 5.60 6.95 8.98 1.5x of Base Case 4.32 5.26 6.68 2.5x of Base Case 2.50 2.93 3.57 Rating Sensitivity Base Case Loss Proxy AAAsf AAAsf AAAsf 1.5x of Base Case AAsf AAAsf AAAsf 2.5x of Base Case BBBsf Asf Asf Class B Notes (x) Base Case Loss Proxy 4.46 5.50 7.05 1.5x of Base Case 3.43 4.12 5.17 2.5x of Base Case 1.97 2.24 2.66 Rating Sensitivity Base Case Loss Proxy AAsf AAAsf AAAsf 1.5x of Base Case Asf AAsf AAAsf 2.5x of Base Case BBBsf BBBsf BBBsf Class C Notes (x) Base Case Loss Proxy 3.34 4.06 5.14 1.5x of Base Case 2.54 2.99 3.66 2.5x of Base Case 1.43 1.56 1.75 Rating Sensitivity Base Case Loss Proxy Asf AAsf AAAsf 1.5x of Base Case BBBsf Asf Asf 2.5x of Base Case BBsf BBsf BBsf Class D Notes (x) Base Case Loss Proxy 2.66 3.14 3.85 1.5x of Base Case 1.70 1.90 2.20 2.5x of Base Case 0.93 1.02 1.18 Rating Sensitivity Base Case Loss Proxy BBBsf Asf AAsf 1.5x of Base Case BBsf BBBsf BBBsf 2.5x of Base Case Bsf Bsf Bsf Class E Notes (x) Base Case Loss Proxy 1.73 2.04 2.50 1.5x of Base Case 1.10 1.23 1.43 2.5x of Base Case 0.66 0.58 0.62 Rating Sensitivity Base Case Loss Proxy BBsf BBBsf BBBsf 1.5x of Base Case Bsf BBsf BBsf 2.5x of Base Case < Bsf < Bsf < Bsf


Structured Finance

Fitch expects to receive legal opinions confirming the cash flow derived from assets (or any other relevant transaction party, such as a swap counterparty) will not be impaired or diminished. Depending on the legal structure of the transaction, Fitch expects to see opinions addressing, among other things, the isolation of assets from the bankruptcy/insolvency risk of the originator and parent of the issuer (nonconsolidation); the grant of a first-priority, perfected security interest in the assets for the benefit of noteholders; and the tax status of the SPV issuer, either as a tax-neutral entity or, if the SPV issuer is taxable, the nature and amount of such taxes. Additionally, where applicable, Fitch expects opinions to address any ERISA-related risks. Opinions provided may vary for different jurisdictions, and material differences will be noted in the transaction rating report.

Vehicle Titling Auto loans are unique in that the sale or assignment of security interest in the automobiles is governed by certificate of title laws in the applicable states; thus, the sale and assignment of security interests in automobiles are noted on the title certificate for each automobile. Because of the expenses and difficulty involved in re-titling automobiles, the certificates of title are not in most cases amended to reflect the subsequent assignments of the liens on the automobiles to the various parties in the transactions.

When there are material concentrations of automobiles/loans in particular states, Fitch expects that local counsel opinions will be provided to confirm that the issuer and the indenture trustee will have a first-priority, perfected, security interest in the automobiles in those particular states, notwithstanding the fact that the certificate of title is not amended to reflect the indenture trustee’s security interest.

E-Contracts Fitch’s criteria and methodology for E-contracts are fundamentally identical to those for traditional auto loan contracts from a legal perspective. E-contracts are originated through the Internet and executed electronically. As with standard-originated, paper auto loan contracts included in a securitized pool of loans, Fitch expects E-contracts to be properly perfected under the relevant provisions of the UCC, and Fitch expects to receive legal opinions addressing the first-priority, perfected, security interest in the E-contracts.

Counterparty Analysis The following section highlights counterparty risks that are common in U.S. and Canadian auto loan ABS transactions; however, they should be considered in conjunction with the relevant counterparty risk and global structure finance criterion. For more information on counterparty risk, see Fitch’s Counterparty Criteria (commingling, hedge, account bank and investments) and Global Structured Finance Rating Criteria. In the case where a servicer provides servicing advances, Fitch expects either the servicer providing the advancing or the trustee acting as a backup to funding advances to be a highly rated institution, in accordance with the provisions of Fitch’s Counterparty Criteria.


Structured Finance

Originator/Seller/Servicer Operational Analysis The originator/seller/servicer, as transaction participants, can affect the performance of the underlying assets and, ultimately, an auto loan ABS transaction. Fitch's ABS analysts review the operational processes for each originator/seller/servicer participating in a SF transaction rated by Fitch.

The result of Fitch’s review may lead to adjustments to a transaction's assumptions and/or stresses, application of a rating cap, or cause Fitch to decline to rate a transaction. These may include poor financial or operational strength; inadequate ability or lack of experience in servicing; and lack of or inadequate financial, operational or performance data/information provided by the applicable party. Such adjustments will be discussed in the transaction rating report.

Fitch expects to conduct a review of the sponsor/seller/servicer’s operations, including their originations, underwriting and servicing divisions, combined with a corporate review, prior to assigning ratings for new issuers. Thereafter, Fitch conducts an operational review approximately every one to two years. Fitch’s Financial Institutions and/or Corporate groups collaborate with Fitch’s ABS group for initial and often subsequent operational reviews.

Fitch’s operational analysis focuses on three main factors: • Corporate performance, including operational and financial stability. • The capabilities and quality of the originations and underwriting processes. • The capabilities and quality of the loan servicing operations. • Credit file reviews (underwriting file reviews) — see Appendix 2 on page 35.

Corporate background and performance is a function of management quality and financial strength, as are operational risk management capabilities. Origination and underwriting strength is evaluated by reviewing the originator’s origination and sourcing procedures, underwriting integrity, quality control and risk management within the origination function. A

Auto Loan ABS Basic Legal Structure

Obligors

True Sale of Receivables

True Sale and/or First-Priority PerfectedSecurity Interest in Collateral

XYZ Company (Parent)

XYZ Finance Corp.(Originator/Servicer)

XYZ Receivables Corp.Bankruptcy-Remote SPV

(Seller)

XYZ Auto Owner Trust 2009-A Not Association or PTPTaxable as Corporation

(Issuer)

SecuritizationNotes/Certificates Sold to Investors

ABS Trust Company Delaware(Owner Trustee)

Bank A, N.A.(Indenture Trustee)

First-Priority Perfected SecurityInterest in Collateral

Ownership Certificates

Ownership Certificates


Structured Finance

review of servicing involves loan and title maintenance, billing, customer service, collection and loss mitigation, investor reporting and risk management.

Corporate Overview An understanding of the company’s history, structure, strategic objectives, management experience and funding capabilities is important in the operational review undertaken by Fitch. Fitch investigates the motivation for the securitization transaction, as well as the role of securitization within the originator’s overall funding strategy.

The stability of a company and its capability of meeting business objectives, paired with its ability to deal with change, reflect management strength. Operational factors in the review are: financial condition; legal structure and ownership; management stability, tenure and depth; financial condition; operating history and experience; compliance and corporate governance; competitive position; portfolio characteristics; and competitive growth strategy.

A final component of the analysis is corporate governance. Integrity of the audit process and regulatory compliance (where required) are elements of governance that are pertinent for loan origination, underwriting and servicing. Fitch expects an issuer to comply with all audit processes and regulatory compliance.

Originations Fitch reviews different origination channels used by the originator to understand the impact they may have on the pool credit quality and performance. Originations should be complementary to a seller/servicer’s core competencies and strategic objectives, including maintaining adequate long-term profitability. The origination staff organization, including reporting structure and segmentation, whether by geographic region, target market or product type, is reviewed. The originator’s level of automation and procedures for prospect evaluation are key factors in the evaluation of a seller/servicer’s originations.

Factors and measures considered by Fitch in its review of the loan origination process and strategy include origination and marketing strategies and account scoring and pricing. The agency reviews the data receipt, verification and loan-approval process. In the case of decentralized loan-approval processes, Fitch also reviews the processes employed by the originator to monitor and control underwriting quality.

Fitch reviews the originator’s processes for reviewing and updating scorecard assumptions (to assess whether the assumptions are kept up to date while being relatively stable).

Underwriting The quality and consistency of the underwriting process are critical to the future performance of receivables. Fitch focuses on the strength and integrity of the risk management, underwriting criteria credit approval, exception management and funding processes. Fitch reviews metrics used in scorecard analysis and examines how often the predictive value of the scorecard is tested and the frequency with which the scorecard is refined and updated. The degree of automation as measured by auto approval and decline rates is also reviewed. Ultimate loss is a function of frequency of default and severity of loss. The credit characteristics of the borrower largely dictate the frequency of default, and they are assessed through a combination of credit underwriting models and underwriter judgment.


Structured Finance

Fitch expects to conduct a credit file review during onsite visits with issuers. This is discussed further in Appendix 2 on page 35.

Servicing Fitch reviews the organizational structure of the servicing functions and the ability to accommodate anticipated growth. If the originator outsources significant portions of the servicing to a third party, a review is also carried out with this party. Servicing functions that Fitch focuses on include the following: • Title maintenance. • Account maintenance. • Customer service. • Billing and payment processing. • Collections. • Repossession and disposition. • Chargeoffs and recoveries. • Investor reporting and remittance. • Staffing and training. • A servicer’s business systems, among other factors. • Backup servicers or replacement of servicers.

If the servicer is non-investment grade or servicing a more volatile portfolio requiring greater intensity and specialization, Fitch assesses the transferability of portfolio servicing. This will involve determining trustee responsibilities; depth of the third-party market; compatibility of the servicing platform and reporting; potential costs associated with the transfer, priority and adequacy of the servicing fee within the securitization waterfall. If the financial condition of the seller/servicer is particularly weak, a formal, paid backup servicing arrangement may be put in place by an issuer, under which parallel processing is performed. Fitch also reviews any backup servicing agreement in place and/or the ability of a servicer to be replaced, and the transferring of servicing to a backup servicer.

Commingling and Payment Interruption Risk In auto loan ABS transactions, the servicer often collects debtor payments into its own bank accounts before transferring such funds to the accounts of the SPV. This provides operational simplicity to the servicer. The commingling period refers to the number of days that collections are held by the originator before being transferred to the SPV. Fitch’s commingling criteria are detailed in its Counterparty Criteria.

In the consideration of commingling/payment interruption risk, Fitch applies a one-month payment period in auto loan ABS transactions given the depth of the auto loan ABS servicer market, which has an adequate market of potential servicers and third-party servicers present, who can service a portfolio upon transfer. Additionally, it is assumed, and has occurred in limited instances, that a servicing transfer could take place over a short period, within one month, and a new servicer can efficiently assume a portfolio and service it.

In instances where the servicer does not meet the minimum rating thresholds outlined in the Counterparty Criteria, Fitch would expect a liquidity support provision to mitigate the risk of interruptions to the payment of interest for any of the rated notes. An example of such a liquidity support provision could include a cash reserve account sized to cover at least one month of scheduled interest and servicing cost distributions.


Structured Finance

Performance Analytics The ongoing performance analysis of transactions forms an essential part of Fitch’s rating process. Fitch conducts a periodic analysis of the bond and collateral performance after issuance. Fitch’s surveillance process is employed to monitor performance and facilitate reviews of transactions in a timely manner. If data collected by Fitch show that transaction performance begins to fall outside of Fitch’s initial expectations or unforeseen events occur (e.g. counterparty related issues) that could impact a transaction’s performance, a review is conducted that could lead to a rating action, as described below. Although monitored monthly, each transaction must be presented to a rating committee at least once a year.

Fitch expects to receive servicer reports from servicers/issuers/trustees monthly, detailing note paydown, available credit support and certain asset performance and portfolio characteristics. Transaction-specific performance on public transactions is posted on Fitch’s surveillance website. Metrics such as bond amortization, delinquencies, gross defaults, recoveries, net losses and CE levels are tracked and available to investors.

Fitch will contact the servicer/issuer if additional detail is needed regarding performance changes within the transaction. Additional information requests may include updated collateral stratifications (such as current WA LTV ratios and current WA FICO scores) and performance statistics on certain sectors or collateral types.

Screener Process Using the data gathered from the monthly servicer reports and aggregated in Fitch’s internal database, the surveillance analyst evaluates collateral characteristics and bond performance metrics while considering microeconomic and macroeconomic issues affecting the issuer and overall health of the auto industry. These collateral characteristics and bond performance metrics are then analyzed and compared with historical levels/metrics and Fitch’s internal assumptions and expectations, and then compared with overall industry trends and peers on an ongoing basis. Fitch will contact the servicer/issuer if additional detail is needed regarding performance changes within the transaction.

Fitch analysts aggregate various high-level transaction information and performance metrics for each active transaction. The results are discussed monthly as part of Fitch’s auto ABS screener process. This comparative analytical approach enables Fitch analysts to differentiate between performance trends observed across the sector and those affecting select or single issuers. Key information presented as part of an auto ABS transaction screening includes: • Current and initial rating, including Rating Watch (if any) and Rating Outlook status. • Current and previous CE. • Number of months outstanding and expected number of months remaining. • Transaction seasoning. • Current pool factor. • Various delinquency buckets. • Initial base case CNL proxy, current CNL and pool factor extrapolated CNL.


Structured Finance

Material changes in key macroeconomic factors, such as unemployment or fuel prices, are assessed to determine any potential impact on the ABS transactions.

The above data are run through various internal algorithms that identify classes of transactions that may be potentially upgraded or downgraded based on these metrics. Transactions with performance falling outside of Fitch’s initial expectations, those building significant levels of CE as they de-lever or transactions approaching one year since issuance or their prior rating review will be reviewed in a rating committee.

Transaction Review Process During the rating review, a surveillance analyst compares actual losses to date incurred by the transaction with the initial base case loss derived at the closing of the transaction. Additionally, other performance-related measures will be examined, such as amortization profiles, portfolio delinquencies, recoveries, losses and the amount of CE built to date. Based on actual losses incurred to date, the surveillance analyst may revise the initial base case loss assumption to account for such losses utilizing a straight-line, pool factor or loss curve projection, as displayed in the table to the right.

The straight-line projection, usually the most conservative projection utilized by Fitch, is calculated by dividing the total CNL to date by the number of payment periods to date. The result is then added to each subsequent period for the expected remaining WA maturity of the transaction. Limitations of this method are that it assumes losses occur at the same rate every payment period when, in fact, actual auto loan ABS losses tend to taper off as the transaction matures. Because this approach is overly punitive, it is usually viewed by Fitch as the maximum CNL projection.

The pool factor projection is calculated by dividing total CNLs to date by the collateral amortization (1 – current pool factor) experienced to date. This projection is most often used by Fitch in transaction reviews, given that it takes into account the transaction’s actual losses to date and applies them to only the remaining portfolio.

Example — Revised Base Case Loss Assumptions Assumptions (A) Initial Base Case Loss (%) 1.50 (B) Net Losses to Date (%) 2.00 (C) Current Pool Factor (%) 30.00 (D) Original Loss Curve at Month 36 (%) 85.00 (E) Current Payment Period (Months) 36 (F) Expected WAM (Months) 60 (G) Remaining Periods (Months) 24 Calculations Straight-Line Projection (%) B + (B/E x G) 3.33 Pool Factor Projection (%) B/(1 − C) 2.85 Loss Curve Projection (%) B/D 2.35


Structured Finance

The loss curve projection is calculated by dividing total CNLs incurred to date by the initial loss curve created at transaction close (see the Loss Curve Example chart above). Limitations of this method are driven by the fact that initial loss timing expectations may not be consistent with current/actual loss timing performance. For a representation of all of Fitch’s loss projections, see the table on page 29.

Surveillance analysts review aforementioned performance-related measures for the transaction in conjunction with the loss curve projections to determine the base case loss proxy for the transaction review. Once a base case loss proxy is determined by the surveillance analyst, the cash flow model is updated and re-run utilizing the updated assumptions (including the updated asset/liability structure). In certain situations, cash flow modeling may not be completed for a subsequent review. For example, in a single class structure with a ‘AAAsf’ rating, if losses are tracking significantly inside expectations and CE has increased since close, Fitch will not incorporate cash flow modeling in its analysis. Another example when Fitch may not incorporate cash flow modeling is in a revolving transaction that has not begun its amortization period, and where credit enhancement levels, portfolio characteristics and performance metrics are materially unchanged from the initial rating. Fitch will complete cash flow modeling for any potential rating actions (at the time an upgrade/downgrade to the rating is recommended). The cash flow model used for surveillance is the same as that used in the initial rating process. The analysis and recommendations are then presented by the surveillance analyst to a rating committee consisting of a required quorum.


Structured Finance

Appendix 1 Additional Adjustments to Derived Loss Proxy

Loss Severity Adjustments Loss severity or, conversely, recovery rates on auto loans are influenced by many factors, as stated in the Recovery Rates section on page 14. For cases in which Fitch expects loss severity to be substantially different from historical experience, loss expectations are adjusted to account for the differential. When provided with gross loss static pool data, Fitch adjusts the assumed recovery rate on defaulted collateral to a level commensurate with its expectations for the term of the securitization. For example, a scenario incorporating a more severe recovery rate stress (i.e. using a lower recovery rate) will result in a higher CNL expectation. In circumstances where Fitch deems a loss severity adjustment necessary but receives only net loss static pool data, the following steps are taken: • Extrapolated net losses are converted to gross losses, utilizing recovery rates for the periods

associated with the data. • Gross losses are converted back to net losses, utilizing the adjusted or stressed recovery

rate assumed to be appropriate for the term of the securitization, as detailed in the table on below, which displays an example of a typical loss severity adjustment applied by stressing historical recovery rates. Going back to 2003, an extrapolated CNL proxy of 1.05% is derived. The historical recovery rate ranges from 45%−52% for those years. Fitch derives the gross losses and assumes a base case recovery rate of 40%, approximately a 10%−20% stress. This results in a higher, final loss proxy of 1.23%, approximately 20% higher than the pre-adjusted loss proxy.

Macroeconomic Adjustments Losses on U.S. and Canadian auto loan ABS and other consumer ABS asset classes have historically displayed considerable correlation with the state of the U.S. labor market, including the unemployment rate and new jobless claims. As the new jobless claims increase, so do loss levels on an auto loan ABS portfolio.

Fitch will apply an adjustment to the loss expectation to account for material, unexpected changes in near-term unemployment levels, which may vary beyond normal historical levels (for example, unemployment levels of 9%–10% in 2008−2010) if deemed appropriate. Fitch considers unemployment adjustments in cases where: • Near-term unemployment expectations are materially different than the current level.

Loss Severity Adjustment An Example (%)

Static Pool CNL Vintage 2003 2004 2005 2006 2007 2008 2009 2010

Avg. Extra-

polation Extrapolated CNL Loss Proxy 1.05 0.98 0.95 1.06 1.00 1.02 1.20 1.16 1.05 Recovery Rate on Vintage 51.00 49.00 45.00 47.00 52.00 50.00 48.00 46.00 Derived Gross Losses 2.14 1.92 1.73 2.00 2.08 2.04 2.31 2.15

Assumed Recovery Rate 40.00 40.00 40.00 40.00 40.00 40.00 40.00 40.00

Avg. Adjusted

Extra-polation

Adjusted CNL Loss Proxy 1.29 1.15 1.04 1.20 1.25 1.22 1.38 1.29 1.23


Structured Finance

• There are fluctuations in the unemployment rate that are expected to be in place for a period that could materially impact securitization performance.

Where deemed appropriate, Fitch applies an unemployment adjustment by adding incremental losses to the base case expected losses derived as described above. This type of adjustment is sized to equal the expected change in the unemployment rate, based on Fitch’s internal projections. This type of adjustment will be noted in the transaction rating report.

The example to the right is one of the methods that Fitch employs to apply an unemployment adjustment. For example, if the U.S. unemployment rate is 10% and expected to rise to 11% in the near term, Fitch would increase the loss expectation for ABC Securitization Trust by 10%, or to 1.65% from 1.50%.

Adjustments to Account for Pool Characteristics Fitch applies additional adjustments to the loss proxy in situations where it feels certain characteristics/components of the collateral pool will perform materially different outside of historical experience. Common characteristics that may be adjusted include the following: • Geographic concentrations in states or regions undergoing notable economic stress. • Concentrations in vehicle segments (e.g. trucks/SUVs/cars) or individual models. • Material migrations in the characteristics of the borrower (e.g. LTV ratios and FICO score).

Fitch expects to receive all static pool data for each characteristic of the pool that may impact the securitization’s performance. If these data are not provided to Fitch or are not available, the loss expectation may be adjusted to account for the potential risk posed by the related collateral characteristic. This type of adjustment will be noted in the transaction rating report.

Geographic Concentrations Obligors in certain U.S. or Canadian states/provinces or regions may represent a disproportionately larger amount of the pool of loans securitized. This can lead to higher pool losses if this region experiences weaker performance and produces higher losses. If this is the case, and Fitch cannot evaluate performance due to the absence of static pool loss data segmented by region, Fitch will adjust the loss expectation in the following manner to account for this risk: • Determine a basis for quantifying the expected performance differential through available

performance (state, provincial or regional managed portfolio performance data) and/or general economic statistics (i.e. state, provincial or regional unemployment rates).

• Apply an additional stress to the expected loss based on the results of the above analysis.

This adjustment is assessed only to the portion of the geographic concentration in the pool that exceeds historical norms for the managed portfolio. Any such adjustment will be described in the transaction rating report.

Unemployment Adjustment An Example (%) Current U.S. Unemployment 10.00 Fitch Expected Unemployment 11.00 Incremental Change (current vs. expected change) 10.00 Unadjusted Base Case Loss Proxy 1.50 Unemployment Adjustment 10.00 Unemployment Adjusted Base Case Loss Expectation 1.65


Structured Finance

Appendix 2 – Issuer Onsite Visit Agenda/Discussion Topics Corporate Overview 1. Company History, Business Lines, Organizational Structure. 2. Ownership, Shareholders (External/Management/Private Equity), Mergers and

Acquisitions. 3. Management Teams and Structure, Staff Experience, Tenure and Biographies. 4. Marketing Strategy, Competitive Landscape and Market Position:

• Future Portfolio Strategies including Non-Prime Volumes and Projections/Growth. • Subprime Market Sector Definition.

5. Sources and Uses of Financing including Securitization Funding. For each of Originations, Underwriting, Servicing and Collections sections listed below, first describe the following and then address the specific points: • Department Structure and Organizational Chart. • Management Teams and Staff Experience, Development/Training and Turnover. • Technology/Tracking Systems to Support Each Functional Team and

Scalability/Maintenance. • Quality Controls and Assurances, Audits/Reporting/Fraud Prevention (Both Internal and

External). Originations and Underwriting Overview A. Originations 1. Current and Future Portfolio Strategies and Goals:

• Acquisition and Growth Targets/Prospects. • Origination Channels and Products. • Target Market/Consumers.

2. Separation of Originations and Underwriting Functions. B. Dealer Networks 1. Dealer Channel – Strategies and Growth Prospects:

• Regional Structure and Markets. • Dealer Penetration Rates. • Major Competitors and Market Trends.

2. Dealer Approval and Relations: • Approval Rates. • Dealer Composition. • Dealers Network – Number of Dealers including Additions/Terminations. • Franchised versus Nonfranchised/Independent Dealers:

o Dealer Underwriting Criteria (Franchised versus Nonfranchised/Independent Dealers).

o Other Criteria: Owner/Location/History/Inventory/Floorplan Performance. o Historical Dealer Composition (Franchised vs. Nonfranchised/Independent

Dealers). 3. Dealer Performance, Reporting and Controls:

• Dealer Profitability and Origination Loan Performance Metrics – Delinquencies/Losses by Dealer.

• Fraud Cases (Count/Impact/Losses).


Structured Finance

• Fraud Prevention Processes and Procedures. • Bankruptcies (Count/Impact/Losses).

4. Dealer Markups: • Amounts/Methodology/Impact on Contract and Consumer/Monitoring. • Dealer Buy-Rates/Markups: Derivation Thereof and Maximums/Minimums. • Discuss Impact on Contract/Portfolio Performance. • Regulatory Oversight.

5. Dealer Repurchases (including breach of R&W): • Repurchase Levels and Reasons for Repurchases. • Enhancements Implemented Related to Repurchase Levels.

6. Dealer Monitoring/Inspections/Reporting, Findings and Enhancements: • Audits/Inspections. • Findings (including Repurchases/Fraud/R&W Breaches). • Enhancements/Controls.

C. Underwriting 1. Overview of Underwriting Criteria. 2. Underwriting and Credit Systems and Credit Utilized. 3. Personnel – Remuneration and Turnover Rates. 4. Dealer Contract Data Verification Policies and Controls. 5. Credit Approval Authorities. 6. Credit Scoring Systems:

• Scorecard Generation. • Frequency of Scorecard Review/Updates. • Scorecard Testing/Results and Enhancements. • Credit Scoring System Inputs – Key Inputs/Drivers/Weightings.

7. FICO and Internal Credit Scoring Analysis/System: • Which version of FICO score is being utilized? • Credit Bureaus Chosen and Their Performance. • Credit Tier Definitions/Cutoffs.

8. Loan-To-Values: • Provide Actual LTV Formula. • LTV Calculation thereof and Industry Vehicle Values Sources. • New versus Used Vehicle Advance Rates. • Add-Ons and Other Products Financed (Taxes/Insurance/Warranties etc.). • LTV Limits by Credit Tiering or Other Metrics.

9. Income Verification: • Criteria Guidelines Related to Income Verification. • Percent of Total Portfolio Income Verified or Not by Criteria (e.g. by Credit Tier). • Credit Tier/FICO Drivers. • Credit Analyst Review Process and Procedures. • Which Applications Do Not Have Income Verified During Process?

10. Employment Verification Criteria and Policies. 11. Other Tools: Lexus/Nexus; Criminal; Rent/Owner Evictions/Social Media. 12. Application Decisioning:

• Decisioning Criteria: o Auto Decisioning versus Analyst Review. o Auto Approvals versus Decline Criteria.


Structured Finance

• Portfolio Ratios and Stats of Auto Approval versus Declines versus Analyst Decisioning.

13. Exceptions Policies and Procedures: • Types/Drivers of Exceptions. • High-Side and Low-Side Overrides and Portfolio Ratios/Stats. • Authority to Process Exceptions and Procedures. • Number/$/% of Total Approved/Funded. • Tracking and Performance of These Contracts.

14. Underwriting Outsourced: • Outsourced Functions/Areas. • Third-Party Underwriters. • Monitoring and Performance.

15. Funding Process: • Policies and Procedures for Funding an Approved Contract. • Pre-Funding Contract Review and Data Verification Parameters Prior to Funding. • Post-Funding Contract Review and Data Verification.

16. Underwriting Validation and Tests: • Quality Assurance: Criteria/Documentation Requirements and Data Accuracy. • Loan Review: Originations versus Established Policies. • Internal Audits: Periodic Internal Control Review of Processes including

Originations/Reporting. 17. Consumer Welcome Call Guidelines and Contract/Obligor Data Confirmed. 18. Rating Agency Sample Credit File Review. Servicing Overview 1. Loan Setup. • Contract Conversion from Underwriting System to Servicing System. • Contract Data Review, Verification and Integrity. • Contract File/Documents Control, Storage and Access. 2. Title Recording, Maintenance and Storage, Including Third-Party Vendors/Name. Customer Service 1. Customer Introduction Call and Review/Verification of Contract Data. 2. Customer Service, Correspondence and Dispute Handling. 3. Call Volumes and Customer Service Rep. Coverage Figures/Ratios. 4. Customer Service Systems Utilized (IVR). 5. Response Parameters. 6. Attrition Rates. 7. Internet/Website and Other Communication Mediums (Phone/Text/Social Media). 8. Customer Satisfaction Measurements and Surveys. 9. Customer Complaints. Billing and Payment Processing 1. Billing Procedures. 2. Remittance Channels:

• Types: Lockbox/check/cash; ACH; Online/Website; Agent Desktop; IVR (automated phone); Recurring; Western Union/MoneyGram.

• Lockbox: Provide Name of Bank. • Other Mediums (/Military/Cellular/Text/Social Media).


Structured Finance

3. Exceptions – Policies, Levels of Authority and Processing Procedures. 4. Account Reconciliation. Collections and Asset Management 1. Collection Strategies, Process and Timelines (Delinquency Buckets and Repossession

Periods). 2. Servicing Systems Utilized. 3. Personnel – Remuneration and Turnover Rates. 4. Servicing Costs:

• Total Costs to Service Portfolio/Securitizations. • Revenues versus Costs / Profitability

5. Servicing Strategies: • Cradle-to-Grave. • Delinquency/Loss/Risk Based.

6. Servicing/Collection Functions Outsourced to Third Parties: • Functions/Vendors/Locations. • Performance/Monitoring. • Compensation Structures.

7. Loss Mitigation Strategy and Policies. 8. Collections Dialer: Strategies and Processes/Policies. 9. Accounts per Collector Data and Ratios. 10. First Payment Defaults. 11. Repossession Process:

• Asset Management Strategies and Policies. • Repossession Agent Strategy. • Recovery Timelines and Rates. • Recovery Rates by Disposition Channels. • Remarketing Management Process. • Disposition Channels and Data. • Dealer Channel. • Wholesale Auctions. • Retail Channels. • Internal Channels (Auctions/Websites/Other).

12. Bankruptcies. 13. Skip Tracing. 14. Charge-Off Policy. 15. Representations and Warranties:

• Dealer-Related Issues. • Number/Percent/Figures.

16. Insurance. 17. Portfolio Monitoring and Reporting Procedures and Policies. 18. Functions, Reporting and Monitoring. 19. Risk Reporting:

• Monitoring of Performance (Delinquencies/Losses), Credit Quality and Exceptions to Servicing Policies.

20. Compliance Governance: • Quality Controls. • Servicing Compliance. • Compliance Control Systems.


Structured Finance

• Compliance Tracking Processes and Systems. • Call Miner Strategies and Programs. • Internal Annual Compliance Servicing Statements. • Accounting Firms’ Servicing Annual Compliance Review and Statements.

21. Servicing Compliance with Regulatory Rules. 22. Rating Agency Sample Collection File Review sample review of at least 10

underwriting/servicing contract files (see Credit File Review at the end of this Appendix). Technology 1. Department Technology and Systems. 2. System Controls and Monitoring. 3. Data Verification Systems, Procedures and Policies. 4. Capacity and Utilization. 5. Quality Assurance and Control Systems. 6. Systems Auditing, Monitoring, Testing and Compliance. 7. Contingency, Backup, Disaster and Recovery Planning. Transaction and Investor Reporting and Remittance 1. Reporting Procedures and Processes. 2. Remittance Procedures and Processes. 3. Internal Controls and Reporting Servicing Report Sample. Regulatory Oversight, Reporting and Investigations 1. Agreed Upon Procedures:

• Reviews, Findings and Enhancements. • Provide AUP from last transaction.

2. Representations and Warranties: • Typical R&W Breaches. • Historical Levels of Related Repurchases/Fraud/Performance of Related Contracts.

3. Offering Documents’ Rule 193 of the S.A. Reviews and Findings: • Procedures and Policies. • Sample Criteria: File Number; Items Reviewed and Number; Total Aggregate Data

Points Checked; Receivables. • Findings and Explanations Thereof, and Overall Conclusions.

4. Exceptions to Underwriting Guidelines: • Number of Exceptions; Explanations; Total Amount Impacted or Statistics. • Third Party Acquisition Data: Entity/Amount/%/Deviations from Internal Criteria.

5. Internal Auditing/Reviews/Investigations: • Frequency of Audits. • Audit Process, Policies and Procedures. • Departments Audited and Parties Responsible. • Findings and Enhancements/Changes Implemented.

6. Accounting Firms’ Audits, Reviews and Findings: • Name of Auditor. • Annual Auditing Compliance Reviews and Letters/Servicer Certificates

(Portfolio/Deals). • Types of Reviews (Periodic/Annual). • Findings and Internal Changes/Enhancement Therefrom.

7. Regulatory Body Oversight: Investigations/Findings/Actions and Enhancements:


Structured Finance

• Securities and Exchange Commission. • Department of Justice. • Fair Lending Procedures and Practices. • Consumer Financial Protection Agency. • Audits, Investigations and Findings.

8. Consumer Complaints. 9. Reporting, Reviews, Audits and Monitoring – Internal:

• Audit Departments and Functions. • Types of Audits. • Compliance. • Audit Findings.

Credit File Reviews • Fitch expects to complete a sample contracts file review to observe the origination and

underwriting processes and practices of an originator, as discussed below. This review is expected to take place during an onsite visit or arranged outside of such visit. o Fitch expects to review at least 5–10 loan contract applications to be provided by the

originator for its sample file review. The applications are expected to cover numerous underwriting file applications with differing decisioning outcomes, including randomly selected auto-approvals, auto-declines, manual approvals and manual declines, as well as various types of underwriting exceptions and processes.

o Fitch reviews the original or electronic application file copies to assess the application and decisioning processes and procedures, including data verification and the presence of key fields/inputs utilized during the underwriting decisioning process.

o Fitch may enquire about certain files and decisioning thereof, including exceptions/ inconsistencies to the stated guidelines and policies against the actual underwriting processes employed in the contracts reviewed.

o If any material observations are observed, Fitch may factor these into the rating process if applicable.


Structured Finance

Appendix 3 Pool Collateral Data Requested

The list below details the collateral stratification data that Fitch expects to receive in accordance with this criteria report for every transaction by an issuer.

Historical Performance Data

Fitch expects an issuer to provide a minimum of five years’ worth of performance data to properly analyze historical performance and trends inherent in its portfolio. The following historical performance data list is expected to be provided to Fitch on transactions by an issuer.

Historical Monthly/Quarterly Static Pool Data with Origination Volume

Stratified by: • Credit scoring (FICO and internal credit scores). • LTV ratios. • Loan term (all terms and OT < 60 months/OT > 60 months). • APR Subvented/nonsubvented collateral (incentive related). • Vehicle collateral (new/used, vehicle type and brand, i.e. truck/car/SUV/van, among other types). • Combinations of several of the above-listed items (e.g. credit band/term/LTV).

Including: • Cumulative defaults (gross). • CNLs. • Recovery rates. • Deficiency balance recovery rates. • Loss to liquidation analysis.

Auto Loan ABS Collateral Stratifications Fitch Expects to Receive Internal Credit Scores Loan Term to Maturity FICO Scores Original Loan Terms Amount Financed Remaining Loan Terms Original Amount Financed Extended (60+ Month) Loan Terms Current Amount Financed Loan-to-Value Ratios Loan Balance Year or Period Originated Original Loan Balance Vehicle Characteristics Current Loan Balance New/Used/Certified Annual Percentage Rate Vehicle Make/Brand/Model All APR buckets Vehicle Segment Car/Truck/SUV/Van/CUV/Other Subvented/Nonsubvented (Incentive) Vehicle Vintage Seasoning Geographic Concentrations State Zip Code Debt-to-Income Payment-to-Income Ratio Obligor Home Ownership Obligor Time at Residence Obligor Time at Current Employment Days Delinquent Times 30 Days Delinquent Lifetime


Structured Finance

• Delinquency buckets. • Pool factor.

Historical Monthly/Quarterly Managed Portfolio Summary Data • Principal receivables outstanding. • Number of contracts outstanding. • Delinquencies by bucket. • Delinquency roll rates. • Defaults. • Repossession count. • Liquidation proceeds. • Losses. • Periodic portfolio characteristics (contract rate distribution, credit scores/tier distribution, LTV

distribution, new/used vehicle composition, vehicle year distribution, vehicle make/model composition, vehicle type composition, original term distribution, remaining term distribution, seasoning, original balance, current balance and geographic distribution).

Performance of Prior Securitizations (If Any/Applicable) • Delinquency, default and loss data. • Prepayment speeds (ABS). • Recovery rates.

Transaction Documents, Blacklined Against the Prior Issuance, If Applicable • Transaction documents. • Offering documents. • Legal opinions. • List of representations, warranties and enforcement mechanisms. • Agreed-upon procedures letter.


Structured Finance

Appendix 4 Modeling Analysis Example (Hypothetical Trust XYZ) This addendum gives additional transparency into Fitch’s modeling analysis by providing a detailed example of the analysis completed on a typical auto loan ABS securitization. This analysis steps through structural aspects of a securitization, collateral characteristics and the basis for repline formation, and the stress and sensitivity analysis conducted.

Structure Hypothetical Trust XYZ issues two classes of notes, class A and B that receive principal on a pro-rata basis. CE consists of a reserve account and overcollateralization. The trust includes a noncurable performance trigger that turns the payment distributions to sequential pay if the rolling three-month average annualized net loss rate exceeds 3.00%.

Collateral Characteristics The Hypothetical Trust XYZ collateral pool can be easily segmented into three main subpools based on the origination characteristics and historical loss and prepayment performance. The subpools include: • 72-month term loans with higher losses and loss severity 45% of the pool. • 60-month subvented collateral with lower prepays 30% of the pool. • 60-month market rate collateral with higher prepays 25% of the pool.

Each of the three subpools is broken into replines based on the remaining term of the collateral for a total of 16 replines, as outlined in the table above. • Prepayments are applied to the individual replines, varying between 1.50% for loans with terms of

more than 60 months as prepayment data reflect that there are higher levels of prepayments for these longer term loans and 0.50% for the loans of less than 60 months, since these loans exhibit lower prepayments.

• Loss severity is assumed to be higher at 60%

Hypothetical Trust XYZ Collateral Replines Balance ($) APR (%) OTerm RTerm

Prepay (ABS) (%) Severity (%)

Loss Speed (Four Years) (%) Base Case (%)

Pool 1 25,000,000 12.00 72 12 1.50 60.0

35−35−20−10

1.60 Pool 2 25,000,000 12.00 72 24 1.50 60.0 1.60 Pool 3 25,000,000 12.00 72 36 1.50 60.0 1.60 Pool 4 125,000,000 12.00 72 48 1.50 60.0 1.60 Pool 5 125,000,000 12.00 72 60 1.50 60.0 1.60 Pool 6 125,000,000 12.00 72 72 1.50 60.0 1.60 Pool 7 30,000,000 10.00 60 12 1.50 50.0

45−35−15−5

1.30 Pool 8 30,000,000 10.00 60 24 1.50 50.0 1.30 Pool 9 30,000,000 10.00 60 36 1.50 50.0 1.30 Pool 10 80,000,000 10.00 60 48 1.50 50.0 1.30 Pool 11 80,000,000 10.00 60 60 1.50 50.0 1.30 Pool 12 20,000,000 3.00 60 12 0.50 50.0

45−35−15−5

1.00 Pool 13 20,000,000 3.00 60 24 0.50 50.0 1.00 Pool 14 20,000,000 3.00 60 36 0.50 50.0 1.00 Pool 15 120,000,000 3.00 60 48 0.50 50.0 1.00 Pool 16 120,000,000 3.00 60 60 0.50 50.0 1.00

Total 1,000,000,000 8.80 65.4 49.5 54.5 Derived Base Case Proxy: 1.35


Structured Finance

• for the loans with terms of more than 60 months, as exhibited by historical performance data, while loans with terms of 60 months or less are assigned lower levels of loss severity, at 50%.

• Loss speeds for loans of more than 60 months are assigned back-ended loss curves (35%/35%/20%/10%), since their loss curves will be more spread out over the term of the loans and occur later in their life, while front-ended curves are applied to the loans with terms of 60 months or less, as these loans display more front-ended loss curves.

Break-Even Results Fitch reviews the proposed liability structure, asset replines and base assumptions, such as prepayments, recoveries and timing curves. Then Fitch stresses the losses and delinquencies on a collateral pool until the first dollar of loss is incurred on each of the rated classes. These break-even stress runs provide insight into the amount of principal defaults a structure could withstand, assuming the other collateral assumptions remain stable. The table on the next page provides an overview on the stress results for Hypothetical Trust XYZ.

Hypothetical Trust XYZ (Break-Even Summary Base Scenario)

Break-Even (%) Break-Even Loss

Multiple (x) Trigger Breach Month of Breach Expected Ratings Class A 6.93 5.14 Yes 6 AAAsf Class B 4.48 3.32 Yes 6 Asf

Example - Modeling Assumptions Loss Timing Curve (% per Year) Front 60−30−10 Base 40−35−17−8 Back 20−25−35−20 Prepayments (Subvented/Market Rate) (%) High 0.7/1.8 Base 0.5/1.5 Low 0.0/1.0


Structured Finance

Appendix 5 Cash Flow Modeling Sensitivity Analysis As part of the break-even modeling process, Fitch conducts sensitivity analyses on the cash flow assumptions of a transaction. This process is extremely important, as it allows Fitch to see the relative effect of cash flow assumptions on principal repayment of the securitized notes. These effects can be large and vary between classes. The analysis also highlights the effects of structural triggers and how they come into play under differing scenarios. The tables above and below detail the assumptions and results of the sensitivity analysis for Hypothetical Trust XYZ.

The class A sensitivity analysis resulted in break-even CNL percentages as high as 8.12% and as low as 6.26%, with changes in prepayments having the greatest effect on loss coverage. Loss speed timing had less of an effect on the transaction, with the base curve resulting in the lowest loss coverage for class A and the highest loss coverage for class B.

Furthermore: • The annualized net loss trigger breached in each of the scenarios, in most cases in the middle of

the first year. • The class B back-ended loss timing scenarios had the latest trigger breach, occurring at the end

of the first year or beginning of the second year of performance.

Hypothetical Trust XYZ Break-Even Summary Sensitivity Analysis

Loss Timing Curve Prepays Break-Even (%)

Break-Even Loss Multiple (x) Trigger Breach Month of Breach

Class A Front Base 7.17 5.31 Yes 5 Class A Base Base 6.93 5.14 Yes 6 Class A Back Base 6.99 5.18 Yes 7 Class A Front High 6.26 4.64 Yes 5 Class A Base High 6.28 4.66 Yes 6 Class A Back High 6.33 4.69 Yes 7 Class A Front Low 8.12 6.02 Yes 5 Class A Base Low 7.78 5.77 Yes 6 Class A Back Low 7.84 5.81 Yes 7 Class B Front Base 4.46 3.30 Yes 6 Class B Base Base 4.48 3.32 Yes 6 Class B Back Base 4.45 3.29 Yes 13 Class B Front High 4.04 3.00 Yes 6 Class B Base High 4.10 3.04 Yes 7 Class B Back High 4.00 2.97 Yes 14 Class B Front Low 5.16 3.82 Yes 6 Class B Base Low 5.18 3.84 Yes 6 Class B Back Low 5.15 3.82 Yes 12

Average Break-Even % by Variable and Class Class A Class B Prepay High 6.29 4.05 Base 7.03 4.46 Low 7.91 5.16

Timing Front 7.19 4.55 Base 7.00 4.59 Back 7.05 4.53


Structured Finance

ALL FITCH CREDIT RATINGS ARE SUBJECT TO CERTAIN LIMITATIONS AND DISCLAIMERS PLEASE READ THESE LIMITATIONS AND DISCLAIMERS BY FOLLOWING THIS LINK: HTTPS://FITCHRATINGS.COM/UNDERSTANDINGCREDITRATINGS. IN ADDITION, RATING DEFINITIONS AND THE TERMS OF USE OF SUCH RATINGS ARE AVAILABLE ON THE AGENCY'S PUBLIC WEB SITE AT WWW.FITCHRATINGS.COM. PUBLISHED RATINGS, CRITERIA, AND METHODOLOGIES ARE AVAILABLE FROM THIS SITE AT ALL TIMES. FITCH'S CODE OF CONDUCT, CONFIDENTIALITY, CONFLICTS OF INTEREST, AFFILIATE FIREWALL, COMPLIANCE, AND OTHER RELEVANT POLICIES AND PROCEDURES ARE ALSO AVAILABLE FROM THE CODE OF CONDUCT SECTION OF THIS SITE. FITCH MAY HAVE PROVIDED ANOTHER PERMISSIBLE SERVICE TO THE RATED ENTITY OR ITS RELATED THIRD PARTIES. DETAILS OF THIS SERVICE FOR RATINGS FOR WHICH THE LEAD ANALYST IS BASED IN AN EU-REGISTERED ENTITY CAN BE FOUND ON THE ENTITY SUMMARY PAGE FOR THIS ISSUER ON THE FITCH WEBSITE. Copyright © 2016 by Fitch Ratings, Inc., Fitch Ratings Ltd. and its subsidiaries. 33 Whitehall Street, NY, NY 10004. Telephone: 1-800-753-4824, (212) 908-0500. Fax: (212) 480-4435. Reproduction or retransmission in whole or in part is prohibited except by permission. All rights reserved. In issuing and maintaining its ratings and in making other reports (including forecast information), Fitch relies on factual information it receives from issuers and underwriters and from other sources Fitch believes to be credible. Fitch conducts a reasonable investigation of the factual information relied upon by it in accordance with its ratings methodology, and obtains reasonable verification of that information from independent sources, to the extent such sources are available for a given security or in a given jurisdiction. The manner of Fitch’s factual investigation and the scope of the third-party verification it obtains will vary depending on the nature of the rated security and its issuer, the requirements and practices in the jurisdiction in which the rated security is offered and sold and/or the issuer is located, the availability and nature of relevant public information, access to the management of the issuer and its advisers, the availability of pre-existing third-party verifications such as audit reports, agreed-upon procedures letters, appraisals, actuarial reports, engineering reports, legal opinions and other reports provided by third parties, the availability of independent and competent third-party verification sources with respect to the particular security or in the particular jurisdiction of the issuer, and a variety of other factors. Users of Fitch’s ratings and reports should understand that neither an enhanced factual investigation nor any third-party verification can ensure that all of the information Fitch relies on in connection with a rating or a report will be accurate and complete. Ultimately, the issuer and its advisers are responsible for the accuracy of the information they provide to Fitch and to the market in offering documents and other reports. In issuing its ratings and its reports, Fitch must rely on the work of experts, including independent auditors with respect to financial statements and attorneys with respect to legal and tax matters. Further, ratings and forecasts of financial and other information are inherently forward-looking and embody assumptions and predictions about future events that by their nature cannot be verified as facts. As a result, despite any verification of current facts, ratings and forecasts can be affected by future events or conditions that were not anticipated at the time a rating or forecast was issued or affirmed. The information in this report is provided “as is” without any representation or warranty of any kind, and Fitch does not represent or warrant that the report or any of its contents will meet any of the requirements of a recipient of the report. A Fitch rating is an opinion as to the creditworthiness of a security. This opinion and reports made by Fitch are based on established criteria and methodologies that Fitch is continuously evaluating and updating. Therefore, ratings and reports are the collective work product of Fitch and no individual, or group of individuals, is solely responsible for a rating or a report. The rating does not address the risk of loss due to risks other than credit risk, unless such risk is specifically mentioned. Fitch is not engaged in the offer or sale of any security. All Fitch reports have shared authorship. Individuals identified in a Fitch report were involved in, but are not solely responsible for, the opinions stated therein. The individuals are named for contact purposes only. A report providing a Fitch rating is neither a prospectus nor a substitute for the information assembled, verified and presented to investors by the issuer and its agents in connection with the sale of the securities. Ratings may be changed or withdrawn at any time for any reason in the sole discretion of Fitch. Fitch does not provide investment advice of any sort. Ratings are not a recommendation to buy, sell, or hold any security. Ratings do not comment on the adequacy of market price, the suitability of any security for a particular investor, or the tax-exempt nature or taxability of payments made in respect to any security. Fitch receives fees from issuers, insurers, guarantors, other obligors, and underwriters for rating securities. Such fees generally vary from US$1,000 to US$750,000 (or the applicable currency equivalent) per issue. In certain cases, Fitch will rate all or a number of issues issued by a particular issuer, or insured or guaranteed by a particular insurer or guarantor, for a single annual fee. Such fees are expected to vary from US$10,000 to US$1,500,000 (or the applicable currency equivalent). The assignment, publication, or dissemination of a rating by Fitch shall not constitute a consent by Fitch to use its name as an expert in connection with any registration statement filed under the United States securities laws, the Financial Services and Markets Act of 2000 of the United Kingdom, or the securities laws of any particular jurisdiction. Due to the relative efficiency of electronic publishing and distribution, Fitch research may be available to electronic subscribers up to three days earlier than to print subscribers. For Australia, New Zealand, Taiwan and South Korea only: Fitch Australia Pty Ltd holds an Australian financial services license (AFS license no. 337123) which authorizes it to provide credit ratings to wholesale clients only. Credit ratings information published by Fitch is not intended to be used by persons who are retail clients within the meaning of the Corporations Act 2001.

https://www.fitchratings.com/creditdesk/public/ratings_defintions/index.cfm?rd_file=intro%22%20%5Cl%20%22lmt_usage

http://www.fitchratings.com/