GCAT 2021-NQM3

Presale:

GCAT 2021-NQM3June 16, 2021

Preliminary Ratings

ClassPreliminaryrating(i) Class type

Initial interestrate (%)(ii)

Preliminary amount($)

Credit enhancement(%)(iii)

A-1 AAA (sf) Senior Fixed 213,413,000 26.55

A-1X AAA (sf) Senior IO Fixed 213,413,000(iv) N/A

A-2 AA (sf) Senior Fixed 21,356,000 19.20

A-3 A+ (sf) Senior Fixed 19,757,000 12.40

M-1 BBB (sf) Mezzanine Fixed 19,468,000 5.70

B-1 BB (sf) Subordinate Fixed 7,409,000 3.15

B-2 B (sf) Subordinate Fixed 5,230,000 1.35

B-3 NR Subordinate Net WAC 3,922,843 0.00

A-IO-S NR Excess servicing (v) Notional(vi) N/A

X NR Monthly excess cashflow

(vii) Notional(vi) N/A

R NR Residual N/A N/A N/A

(i)This presale report is based on information received as of June 16, 2021. The ratings shown are preliminary. Subsequent information mayresult in the assignment of final ratings that differ from the preliminary ratings. Accordingly, the preliminary ratings should not be construed asevidence of final ratings. This report does not constitute a recommendation to buy, hold, or sell securities. The preliminary ratings address ourexpectation for the ultimate payment of interest and principal. (ii)Interest can be deferred on the classes. Fixed coupons on class A-1, A-1X,A-2, A-3, M-1, B-1, and B-2 certificates are subject to the pool's net WAC. Interest on class B-3 certificates equals pool's net WAC. (iii)Thiscredit enhancement is solely from subordination. Excess spread also provides credit enhancement. (iv)Class A-1X will have a notional amountequal to the lesser of (a) the balance of class A-1 immediately prior to such distribution date and (b) the notional amount set forth on aschedule for the related accrual period. After the 36th distribution date, the notional amount of the A-1X certificates will be zero. (v)Excessservicing strip and reimbursable advance. (vi)The notional amount equals the aggregate principal balance of the loans. (vii)Certain excessamounts from the pool's net WAC over classes with fixed-rate coupons calculated per the pooling and servicing agreement. WAC--Weightedaverage coupon. IO—Interest-Only. NR--Not rated. N/A--Not applicable.

Profile

Expected closing date June 25, 2021.

Deal cut-off date June 1, 2021.

Distribution date The 25th of each month, or the next business day, beginning July 2021.

Stated maturity date May 25, 2066.

Presale:

GCAT 2021-NQM3June 16, 2021

PRIMARY CREDIT ANALYST

Meghan Benegar

Centennial

+ 1 (303) 721 4658

[email protected]

SECONDARY CONTACT

G C Torres

San Francisco

+ 1 (415) 371 5066

[email protected]

SURVEILLANCE CREDIT ANALYST

Truc T Bui

San Francisco

+ 1 (415) 371 5065

[email protected]

ANALYTICAL MANAGER

Vanessa Purwin

New York

+ 1 (212) 438 0455

[email protected]

www.standardandpoors.com June 16, 2021 1

© S&P Global Ratings. All rights reserved. No reprint or dissemination without S&P Global Ratings' permission. See Terms of Use/Disclaimeron the last page.

2668248

mailto: [email protected]








Profile (cont.)

Certificate balance,including unrated classes

$290.56 million in aggregate.

Collateral types First-lien fixed- and adjustable-rate fully amortizing and interest-only residential mortgageloans primarily secured by single-family residential properties, planned-unit developments,condominiums, condotels, cooperatives, two- to four-family residential properties andmanufactured home properties to both prime and nonprime borrowers. The pool has 516 loans,which are nonqualified, ATR-exempt, or QM – HPML mortgage loans.

Collateral U.S. residential mortgage loans.

Credit enhancement For each class of preliminary rated certificates, subordination in the form of certificates that arelower in payment priority, as well as excess spread that preserves subordination.

ATR--Ability to repay.

Participants

Issuer GCAT 2021-NQM3 Trust.

Sponsor Blue River Mortgage II LLC.

Seller MAT II LLC.

Depositor GCAT NQM Depositor II LLC.

Certificate registrar, trustee, andsecurities administrator

Citibank N.A.

Master servicer Nationstar Mortgage LLC.

Servicers NewRez LLC (d/b/a Shellpoint Mortgage Servicing) and Citadel Servicing Corp.

Servicing administrator Red Creek Asset Management LLC.

Custodian U.S. Bank N.A.

Originators Citadel Servicing Corp. (d/b/a Acra Lending), Arc Home LLC, Change Lending LLC(f/k/a Commerce Home Mortgage LLC), Quontic Bank, and HomeXpress MortgageCorp.

Loan data agent DV01 Inc.

D/b/a--Doing business as. f/k/a--formerly known as.

Primary Originators

Entity By balance (%) Due diligence (%) Originator ranking

Acra Lending 43.5 100 N/A

Arc Home LLC 31.5 100 N/A

Change Lending LLC 11.7 100 N/A

Quontic Bank 8.3 100 N/A

HomeXpress Mortgage Corp. 5.0 100 N/A

N/A--Not applicable.



2668248

Presale: GCAT 2021-NQM3

Servicers

Entity By balance (%)S&P Global Ratings' selectservicer Operation

NewRez LLC (d/b/a Shellpoint MortgageServicing)(i)

56.5 Yes Primary servicer

Citadel Servicing Corp. 43.5 No Primary servicer

Nationstar Mortgage LLC 100.00 Yes Master servicer

(i)As of the cut-off date, approximately 3.3% mortgage loans of the pool balance are currently interim serviced and are scheduled to transfer toShellpoint Mortgage Servicing before the closing date. D/b/a--Doing business as.

Rationale

The preliminary ratings assigned to GCAT 2021-NQM3's mortgage pass-through certificatesreflect our view of:

- The asset pool's collateral composition (see the Collateral Summary section below);

- The transaction's credit enhancement;

- The transaction's associated structural mechanics;

- The transaction's geographic concentration;

- The transaction's representation and warranty (R&W) framework;

- The mortgage aggregator, Blue River Mortgage II LLC (BRM II); and

- The impact the COVID-19 pandemic will likely have on the performance of the mortgageborrowers in the pool (see "Economic Outlook U.S. Q2 2021: Let The Good Times Roll,"published March 24, 2021) and the liquidity available in the transaction.

S&P Global Ratings believes there remains high, albeit moderating, uncertainty about theevolution of the coronavirus pandemic and its economic effects. Vaccine production is ramping upand rollouts are gathering pace around the world. Widespread immunization, which will help pavethe way for a return to more normal levels of social and economic activity, looks to be achievableby most developed economies by the end of the third quarter. However, some emerging marketsmay only be able to achieve widespread immunization by year-end or later. We use theseassumptions about vaccine timing in assessing the economic and credit implications associatedwith the pandemic (see our research here: www.spglobal.com/ratings). As the situation evolves,we will update our assumptions and estimates accordingly.

Environmental, social, and governance (ESG)

Our rating analysis considers a transaction's potential exposure to ESG credit factors. For RMBS,we view the exposure to environmental credit factors as average, social credit factors as aboveaverage, and governance credit factors as below average (see "ESG Industry Report Card:Residential Mortgage-Backed Securities," published March 31, 2021). In our view, thetransaction's exposure to social and geographic credit factors are in line with the sectorbenchmark. For RMBS, we generally consider social credit factors as above average as housing isviewed as one of the most basic human needs, and conduct risk presents a direct social exposure



2668248


for lenders and servicers because regulators are increasingly focused on ensuring fair treatmentof borrowers. Social risk is generally factored into our base-case assumptions for RMBStransactions based on our consideration of the origination platform, the R&W framework and thethird-party due diligence that informs our view of credit underwriting and compliance withapplicable consumer protections.

The transaction's governance risk exposure is higher than our benchmark due to certain limitsrelated to length of originator loan performance history and the financial capacity of partiesproviding the representations and warranties (R&Ws). By applying certain mortgage operationalassessment (MOA) and R&W pool-level adjustment factors to the transaction, we have accountedfor risk related to ESG credit factors. Certain other features also provide mitigants to thetransaction's governance risk exposures including the fact that 100% of the loans in the pool weresubject to a third-party regulatory compliance due diligence review (see the Third-Party DueDiligence section for more detail).

Overview

GCAT 2021-NQM3 is the ninth RMBS transaction rated by S&P Global Ratings under the GCATshelf. Non-qualified mortgage (non-QM) loans make up 61.6% of the pool, while 38.4% are exemptfrom the QM rules because they are either investor property loans or loans originated by QuonticBank (Quontic) or Change Lending LLC (Change Lending), which are community developmentfinancial institutions. Additionally, one loan (0.03% of the pool balance) is a qualified higher pricedmortgage loan (QM-HPML). The seller, MAT II LLC, aggregated the loans from five originators viaBRM II-affiliated parties. The non-zero weighted average original FICO score of the pool is 734,which is slightly lower than the previous two transactions, but higher than those for prior GCATtransactions. The weighted average original combined loan-to-value (CLTV) of the pool is 69.3%,which is in line with the last four GCAT transactions. (see chart 1).

Chart 1



2668248


Noteworthy Features

Loans in forbearance

On March 27, 2020, the Coronavirus Aid, Relief, and Economic Security (CARES) Act enactedCOVID-19 pandemic-related relief for borrowers with government-backed mortgage loans in theform of a temporary forbearance of up to 12 months of scheduled payments. While non-agencyloans do not fall under the CARES Act as it relates to this forbearance, servicers have beengranting forbearance plans to non-agency borrowers as well, typically with some variations tothose of the CARES Act (e.g., time frame, approval requirements, etc.). The updates we made onApril 17, 2020, to our mortgage outlook and corresponding archetypal foreclosure frequency levels(see "Guidance: Methodology And Assumptions For Rating U.S. RMBS Issued 2009 And Later,"published April 17, 2020) account for a portion of borrowers entering COVID-19 pandemic-relatedtemporary forbearance plans and their impact to the overall credit quality of collateralized pools.To the extent a securitization pool exhibits growth levels in forbearance over time beyond thoseotherwise expected, additional adjustments may be applied.

To differentiate the credit quality of securitization pools with varying percentages of loans inactive (or recently completed) forbearance or deferment due to the outbreak of the COVID-19pandemic at the time of issuance, we increased loss coverage levels to account for the potentialincremental risk. Given our current expectations for temporary forbearance or deferment plansand our market outlook, we view the credit quality of a mortgagor on a forbearance or defermentplan as weaker than one with a current loan, but potentially stronger than one with a 30-daydelinquent loan that exhibits payment issues in a normal macroeconomic environment. Our viewconsiders that forbearance or deferment may have been utilized by some borrowers who couldhave otherwise made the payment due, or the forbearance may be related to a temporary furloughor loss of income. The adjustment factors we apply to 30- and 60-day delinquent loans are 2.5xand 5.0x, respectively.

As of June 1, 2021, 43 mortgage loans (approximately 5.9% of the pool balance) were previouslysubject to a forbearance plan of up to nine months and have received a payment deferral due tothe outbreak of the COVID-19 pandemic. These loans have completed their forbearance plans andhave continued to make full monthly payments after their deferrals. As of the cut-off date, thepre-closing deferred amounts had an approximate outstanding balance of $395,057. Theseamounts, which do not accrue interest, are not pledged to the rated certificates and aresubordinated to accrued interest and principal upon liquidation. Therefore, we accounted forthese deferred amounts to determine the loan-to-value (LTV) ratio for the affected loans withrespect to foreclosure frequency, but not loss severity. We applied a forbearance-relatedadjustment factor of 1.50x for these loans. This resulted in an overall adjustment factor of 1.05x atthe pool level.

When deriving these factors, we considered aspects such as the seasoning of the loans andforbearance plans, payment patterns of those loans before and throughout the forbearance plan,the various stages of forbearance or deferment (see table 1), and our general expectations ofadditional forbearance and deferment from now until securitization closing.



2668248


Table 1

Forbearance/Deferment Status

Loan count (no.) % by balance

Never received forbearance or deferment 473 94.1

Completed forbearance plan and/or deferment granted 43 5.9

We will continue to monitor the credit behavior related to forbearance and deferrals as thesituation evolves and more performance information becomes available, and we may adjust ourloss coverage levels accordingly, which could affect the ratings. For instance, if we were to changethe pool-level adjustment related to the portion of the pool that has received COVID-19-relatedforbearance relief to 1.10x (which is more akin to our adjustment factors for 30- to 60-daydelinquent loans) from 1.05x, the preliminary ratings could, in some cases, be approximately onenotch lower. We will also continue to monitor macroeconomic and housing conditions and updateour mortgage market outlook and associated archetypal foreclosure frequencies, as applicable.

A-1X class

The transaction includes class A-1X as a senior interest-only class. On each distribution date, theclass A-1X certificates will have a notional amount equal to the lesser of the class balance of theA-1 certificates and the notional amount set forth on a schedule for the related accrual period.However, after the 36th distribution date, the notional amount of the class A-1X certificates will bezero, at which time we may withdraw our rating. The pass-through rate for the class A-1Xcertificates will equal the lesser of (1) the fixed rate (3.50%) for the related accrual period and (2)the greater of (a) zero and (b) the net WAC rate for the related distribution date minus thepass-through rate for the class A-1 certificates for such accrual period.

Servicing

Shellpoint Mortgage Servicing (Shellpoint) must advance delinquent principal and interest (P&I)on any delinquent mortgage loan it services until it is either greater than 90 days' delinquent(limited P&I advancing) or the P&I advance is deemed unrecoverable by the related servicer. This isa change from prior GCAT transactions where servicers were typically obligated to advance up to180 days of P&I; however, it is similar to the GCAT 2021-NQM1 and GCAT 2021-NQM2transactions. For loans serviced by Citadel Servicing Corp. (Citadel), neither Citadel nor any othertransaction party is required to make P&I advances.

We applied a delinquency stress to test the transaction's liquidity due to the limited nature of P&Iadvancing and because advancing is not required on a loan while it is in a state of temporaryforbearance due to COVID-19-related hardships. We assumed that 35.00% of the cut-off loanbalance would be in forbearance (and noncash flowing) for the first six months of the transaction,with any P&I payments related to this delinquent portion coming back to the transaction after alldefaults have been passed through to the transaction (approximately 144 months). Further, toaccount for the lack of advancing on Citadel loans, in accordance with our criteria, we stressedliquidity to account for delinquent and defaulted loans in our cash flow analysis (see Cash Flowsection below).



2668248


High number of large balance loans

Seventy-eight loans (15.12% by loan count) have balances greater than $1 million in the pool.Together, these loans make up 40.9% of the pool balance. From a FICO perspective, these loanshave a much better credit profile, although from an LTV perspective, these loans have a slightlyworse credit profile than the overall pool. These loans have a weighted average used LTV of 69.6%and a weighted average used FICO of 741 as compared to the overall pool, which has a weightedaverage used FICO of 734 and a weighted average used LTV of 69.1%. As a result, our losscoverage estimates for the high-balance loans is 12.1% at the 'AAA' level, compared to 16.6%'AAA' loss coverage (without any additional pool level factors) for the overall pool. We also appliedour large loan balance analysis per our criteria (as further described in the Large Loan And TailRisk Consideration section) and concluded no additional loss coverage adjustment was needed toaccount for the potential risk posed by large loans.

High concentration of alternative/other documentation loans

The GCAT 2021-NQM3 transaction includes 294 alternative documentation loans (61.6% by poolbalance) for which income was verified using bank statements or profit and loss (P&L) statementsand 145 other documentation loans (20.1% by pool balance) that were underwritten either to aninvestment property business purpose program (128 debt service coverage ratio [DSCR] and fiveloans primarily underwritten to FICO scores and LTV ratios, rather than DSCR) or for which incomewas verified using asset depletion (12 loans). The combined concentration of these two categories(81.6% by pool balance) is in line with the previous GCAT transaction but higher than we have seenin other S&P Global Ratings-rated GCAT transactions. As we see from the table below, thisconcentration shows an increasing trend, to the 70% range for the 2020 transactions and up tothe 80% range for the 2021 transactions, from the low 60% range in the 2019 transactions.Similarly, we see an increase, albeit smaller than the combined alternative/other documentationtype concentration, in the weighted average current CLTV ratio from the mid- to high-60% range.The weighted average current FICO score shows an increasing trend across the GCAT shelf,partially offsetting the additional risk from increased concentrations of the documentation typesmentioned above and weighted average current CLTV ratios.

Table 2

Alternative/Other documentation loan concentration

CollateralCharacteristics(i)

GCAT2021-NQM3

GCAT2021-NQM2

GCAT2021-NQM1

GCAT2020-NQM2

GCAT2020-NQM1

GCAT2019-NQM3

GCAT2019-NQM2

GCAT2019-NQM1

WA current CLTV(%)

68.4 69.5 69.1 69.4 69.0 64.4 63.4 65.2

WA FICO(iii) 735 738 742 734 720 735 731 718

Alternative/bankstatementdocumentation(%)

61.6 67.5 68.9 43.2 45.5 39.2 42.1 52.4

Other/assetdepletion/DSCRdocumentation(%)

20.1 21.1 11.1 27.6 24.8 9.6 21.2 13.0



2668248


Table 2

Alternative/Other documentation loan concentration (cont.)

CollateralCharacteristics(i)

GCAT2021-NQM3

GCAT2021-NQM2

GCAT2021-NQM1

GCAT2020-NQM2

GCAT2020-NQM1

GCAT2019-NQM3

GCAT2019-NQM2

GCAT2019-NQM1

CombinedConcentration ofAlternative/OtherDoc loans.

81.6 88.6 80.0 70.8 70.3 48.8 63.3 65.4

Community development financial institution loans

Approximately 20.0% of the loans in pool were originated by Change Lending and Quontic, whichare designated by the U.S. Department of the Treasury as community development financialinstitutions (CDFIs). Loans originated by a CDFI are exempt from the ability-to-repay (ATR) rules;therefore, all Change Lending- and Quontic-originated loans, including the ones for a primaryresidence, are ATR exempt. Since the lender did not have to determine the related borrower's ATRfor the CDFI loans, these loans could have greater risk than mortgage loans for which the ATRrules applied. Despite this, the CDFI loans' collateral profile is strong, with a weighted averagenon-zero original FICO score of 747 and an average original CLTV of 69.9%. We also reviewed theunderwriting guidelines of the Change Lending- and Quontic-originated loans, confirmed with theissuer that income documentation was received and considered for these loans, and believe ourloan-level LEVELS model accounts for any associated risk posed by the CDFI loans (see theDocumentation Type section below for more detail).

Reduced risk retention

Unless an exemption exists, an ABS securitizer is generally required to retain at least 5.0% (therequired credit risk) of the aggregate credit risk of the assets collateralizing the securities. TheCDFI loans in this transaction fall within the definition of community-focused residentialmortgages and are exempt from U.S. risk retention rules. Consequently, because CDFI loanscomprise approximately 20.0% of the pool, the transaction's required credit risk retention will bereduced to 4.0% from 5.0%.

Optional redemption

An optional redemption can occur at the earlier of the June 2024 (three years after the closingdate) distribution date or when the loans' aggregate stated principal balance is less than or equalto 30.0% as of the cut-off date. This feature differs from the GCAT 2021-NQM2 transaction, whichhad a redemption option for two years after the closing date.

Optional purchase of delinquent loans

The depositor or its designee, at its option, may purchase any mortgage loan that is 90 or moredays Mortgage Bankers Assn. (MBA) delinquent (including any loans subject to a forbearance planrelated to the COVID-19 pandemic that are 90 or more days MBA delinquent following the end ofthe forbearance period), or any real estate-owned (REO) property acquired in respect of amortgage loan from the issuing trust at the purchase price, which covers principal at par, accrued



2668248


interest, pre-closing deferred amounts, and relevant expenses. The aggregate-stated principalbalance of such mortgage loans and REO properties purchased by the sponsor cannot exceed7.5% of the aggregate stated principal balance of the mortgage loans as of the cut-off date.

Collateral Summary

GCAT 2021-NQM3's assets consist of fixed- and adjustable-rate QM/HPML, non-QM, and ATRexempt loans (some with interest-only periods) secured by first liens. The asset pool consists of516 mortgage loans with a principal balance of approximately $290.56 million as of the cut-offdate.

The collateral pool, from a credit perspective, is weaker than the S&P Global Ratings' archetypalprime pool, but it is generally in line with our expectations of a nonprime residential mortgagepool. The pool's 'AAA' loss coverage requirement was determined to be 20.75%. In our analysis, weconsidered the following mortgage loan characteristics to be weaker:

- A significant number of non-QM loans;

- Alternative income documentation on loans;

- Property-focused investor loans;

- Loan type (adjustable-rate mortgage loans and interest-only term features);

- Loan terms (40-year term loans);

- Property type;

- Occupancy status: second home or investor property;

- Loan purpose: cash-out refinances; and

- Self-employed borrowers.

The mortgage loans consist of adjustable-rate fully amortizing mortgage loans (12.3% of the poolbalance), fixed-rate fully amortizing mortgage loans (87.7%), adjustable-rate interest-only (4.6%)loans, and fixed-rate interest-only loans (6.1%). The weighted average seasoning for the pool isapproximately four months from the loans' first payment date to the cut-off date.

The weighted average used FICO score of 734 is slightly lower than the previous transaction, butgenerally higher than other transactions in the shelf, and the weighted average original CLTV of69.3% is in line with the last three GCAT transactions. This combination overall resulted in aFICO/LTV adjustment factor of 0.96x in our model, which is slightly higher than the previoustransaction, but lower than the GCAT transactions issued in 2020.

Compared to the prior GCAT transaction, GCAT 2021-NQM3 consists of higher property-focusedbusiness purpose loans and fewer cash-out refinance loans and adjustable-rate loans. GCAT2021-NQM3 also has a lower debt-to-income (DTI) ratio than previous transactions. The strengthscompared to prior GCAT transactions are partly offset by a higher concentration of loans withincome verified using alternative methods such as bank statements (see chart 1 above).

The weighted average used FICO score includes certain S&P Global Ratings assumptions (seetable 3 for a breakdown of the pool by borrower FICO score). In the pool, there are 59 loans toforeign borrowers (6.7% of the pool balance), of which 34 are to borrower(s) without a FICO score.We assessed these loans in our credit analysis using a FICO score of 686, which is approximatelythe mortgage pool's average original FICO score minus one standard deviation. For all loans to



2668248


foreign borrowers, we applied a 1.5x multiple to the foreclosure frequencies.

Table 3

Updated Credit Score Statistics

FICO score Current balance (%) No. of loans Average current balance ($000s)

750+ 42.96 195 640.1

725-749 15.30 75 592.9

700-724 18.51 93 578.4

675-699 13.80 98 409.2

650-674 5.76 33 506.8

625-649 2.10 12 509.4

600-624 0.32 3 313.7

575-599 0.90 4 656.2

550-574 0.15 1 430.9

Below 550 0.19 2 269.2

Total 100.0 516 563.1

Mortgage loans backed by properties that are primary residences make up approximately 74.6%of the pool balance. The mortgage loans are secured by first liens on single-family residences(56.0% of the pool balance), planned-unit developments (PUDs; 26.5%), condominiums (8.9%),co-ops (0.3%), two- to four-family homes (7.5%), and other property types such as manufacturedhousing and condotels (1.0%). (See table 4.)

Table 4

Collateral Characteristics(i)

GCAT2021-NQM3

GCAT2021-NQM2

GCAT2021-NQM1

GCAT2020-NQM2

GCAT2020-NQM1

GCAT2019-NQM3

GCAT2019-NQM2

GCAT2019-NQM1

Archetypalpool(ii)

Closing pool balance (mil.$)

290.6 210.0 289.4 226.9 335.4 314.6 402.9 393 N/A

Closing loan count (no.) 516 377 535 353 621 820 1056 876 N/A

Avg. loan balance ($) 563,093 557,062 540,956 642,773 540,140 383,651 381,604 448,672 N/A

WA original CLTV (%) 69.3 69.7 69.7 69.7 69.5 65.4 65.5 66.2 75.0

WA current CLTV (%) 69.0 69.5 69.1 69.4 69.0 64.4 63.4 65.2 75.0

WA FICO(iii) 734 738 742 734 720 735 731 718 725

WA current rate (%) 5.1 5.2 5.7 5.4 5.8 5.7 5.9 6.1 N/A

WA original term (mos.) 362 361 362 373 381 368 368 361 360

WA seasoning (mos.) 4 3 6 5 3 6 9 9 0-6

WA debt-to-income (%) 32.3 31.5 33.3 37.9 36.5 35.9 36.5 37.3 36.0

WA DSCR (non-zero) 1.0 1.1 1.2 1.1 1.1 1.2 1.2 1.2 N/A

Owner occupied (%) 74.6 73.9 73.5 68.0 70.0 62.8 64.2 76.3 100.0

Single-family (includingunattached and attachedPUD) (%)

82.4 84.3 83.2 77.9 80.0 78.5 77.5 72.3 100.0



2668248


Table 4

Collateral Characteristics(i) (cont.)

GCAT2021-NQM3

GCAT2021-NQM2

GCAT2021-NQM1

GCAT2020-NQM2

GCAT2020-NQM1

GCAT2019-NQM3

GCAT2019-NQM2

GCAT2019-NQM1

Archetypalpool(ii)

Adjustable-rate loans (%) 12.3 14.7 24.6 42.1 56.1 85.2 83.5 70.0 0.0

Loans with IO payments(%)

10.8 13.4 13.7 37.1 34.2 10.7 10.0 4.5 0.0

Purchase (%) 60.0 68.0 68.4 32.3 41.9 62.8 57.4 60.8 100.0

Cash-out refinancing (%) 21.5 19.1 25.4 41.2 42.4 28.0 29.7 32.8 0.0

Full documentation (%) 18.4 11.4 20.0 29.2 29.7 51.1 36.7 34.6 100.0

Alternative/bankstatement documentation(%)

61.6 67.5 68.9 43.2 45.5 39.2 42.1 52.4 0.0

Other/assetdepletion/DSCRdocumentation (%)

20.1 21.1 11.1 27.6 24.8 9.6 21.2 13.0 0.0

Investor property % 22.3 23.5 22.2 29.8 28.5 22.5 33.9 20.3 0.0

Self-employed borrowers(%)

75.1 77.5 76.2 58.9 54.0 43.9 53.3 59.7 0.0

Loans with co-borrowers(%)

25.9 30.2 27.9 36.8 33.2 21.3 22.9 26.3 0.0

Loans to borrowers withmultiple mortgages (%)(iv)

1.9 1.0 3.4 6.5 5.5 1.6 3.6 2.4 N/A

Loans to foreignborrowers (%)(foreignnational andnon-permanent residentaliens)

6.7 4.8 2.7 1.1 3.1 3.7 3.8 4.1 0.0

Modified loans (%)(v) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

PCEs (%)(v) 0.5 1.2 0.3 0.5 0.9 0.3 0.5 0.8 0.0

Current (%) (viii) 98.9 99.9 98.1 100.0 100.0 100.0 100.0 100.0 100.0

30+ day delinquent (%) 1.1 0.1 1.9 0.0 0.0 0.0 0.0 0.0 0.0

Length of P&I advancing(mos.)(vi)

3 3 3 6 6 6 6 6 Full

Pool-level adjustments (multiplicative factors)

Geographicconcentration

1.02 1.05 1.06 1.14 1.09 1.22 1.10 1.18 1.00

Mortgage operationalassessment

1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.05 1.00

Representations andwarranties

1.10 1.10 1.10 1.10 1.10 1.10 1.10 1.10 1.00

Other (i.e. loanmodification/PCE/duediligence)

1.01 1.02 1.00 1.01 1.01 1.01 1.01 1.01 1.00

Loans inforbearance/deferredpayements related toCOVID-19

1.05 1.05 1.00 1.15 N/A N/A N/A N/A N/A



2668248


Table 4

Collateral Characteristics(i) (cont.)

GCAT2021-NQM3

GCAT2021-NQM2

GCAT2021-NQM1

GCAT2020-NQM2

GCAT2020-NQM1

GCAT2019-NQM3

GCAT2019-NQM2

GCAT2019-NQM1

Archetypalpool(ii)

Combined pool-leveladjustments(vii)

1.25 1.30 1.22 1.53 1.27 1.41 1.28 1.38 1.00

Loss estimation (viii)

'AAA' loss coverage(%)

20.75 21.65 17.75 28.35 28.30 18.75 20.50 23.50 7.50

'AAA' foreclosurefrequency (%)

42.53 42.22 37.52 54.31 55.89 37.01 43.08 47.34 15.00

'AAA' loss severity (%) 48.79 51.28 47.31 52.20 50.64 50.66 47.59 49.64 50.00

'BBB' loss coverage(%)

6.90 7.20 5.55 9.70 9.15 5.35 6.00 6.75 1.92

'BBB' foreclosurefrequency (%)

23.47 22.95 20.30 30.66 29.41 18.22 21.76 24.11 6.41

'BBB' loss severity (%) 29.40 31.37 27.34 31.64 31.11 29.36 27.57 28.00 30.00

'B' loss coverage (%) 2.35 2.20 1.60 3.15 2.50 1.40 1.50 1.65 0.65

'B' foreclosurefrequency (%)

10.15 9.47 8.36 13.74 10.80 6.64 7.70 8.48 3.25

'B' loss severity (%) 23.15 23.23 19.14 22.93 23.15 5.31 19.48 6.78 20.00

(i)Information as of cut-off date for each transaction. (ii)As defined in our Feb. 22, 2018, criteria article. (iii)FICO scores reflect the most recent scores obtained with certainanalytical assumptions (iv)Limited to borrowers who have multiple mortgage loans or properties included in the securitized pool. (v)Limited to modified and PCE loans consideredin our analysis. (vi)Months of P&I advancing on a delinquent mortgage loan to the extent such advances are deemed recoverable. (vii)Combined pool-level adjustments are theproduct of each pool-level adjustment listed above. (viii) The guidance document published April 17,2020, reflects a revision of our ‘B’ (base-case) projected foreclosurefrequency assumption for an archetypal loan to 3.25% from 2.50%. WA--Weighted average. CLTV--Combined loan-to-value ratio. DSCR--debt service coverage ratio.PUD--planned-unit development. IO--Interest only. PCE--Prior credit event. P&I--Principal and interest. N/A--Not applicable.

Transaction Structure

Chart 2 shows an overview of the transaction's structure.



2668248


The transaction is structured as a true sale of the receivables from the seller, MAT II LLC, to thedepositor, GCAT NQM Depositor II LLC, and a transfer from the depositor to the issuing trust, GCAT2021-NQM3 Trust. The issuing trust transfers the certificates to the depositor, and the depositorsells them to the initial purchaser, who then sells them to third-party investors. The depositorsells the non-offered certificates, as well as the certificates required to be held to satisfy the riskretention rules, to either the sponsor, Blue River Mortgage II, or a majority-owned affiliate.

In rating this transaction, S&P Global Ratings will review the legal matters it believes are relevantto its analysis, as outlined in its criteria.

Strengths And Weaknesses

We believe the following characteristics strengthen the GCAT 2021-NQM3 transaction:

- The mortgage pool generally consists of loans to borrowers with higher FICOs and significanthome equity relative to our archetypes, as demonstrated by the pool's weighted average usedFICO of 734 and original CLTV ratio of 69.3%.

- The third-party due diligence providers--Clayton Services LLC, Consolidated Analytics LLC,



2668248


Digital Risk LLC, Evolve Mortgage Services LLC, and Recovco Mortgage Management LLC (all ofwhich are on our list of reviewed providers)--performed due diligence on 100% of the pool'sloans. Their reviews encompassed credit (underwriting) compliance, property valuations,regulatory compliance, and data quality.

- The class A-1, A-2, and A-3 certificates (the senior classes) benefit from a credit support floorin which no principal is paid to the mezzanine and subordinate classes until the senior classesare retired. Additionally, principal is paid sequentially among the senior classes in periodswhen either the cumulative loss or the delinquency trigger has failed, further protecting themore-senior classes.

We believe the following factors weaken this transaction:

- Income on certain mortgage loans (61.6% of the pool balance) was verified using alternativemethods, such as personal or business bank statements, or certified public accountant(CPA)-prepared P&L statements. We consider income verification using those alternativemethods to be a weaker standard than "full" documentation of income. Consequently, weincreased our loss coverages for these loans by applying an adjustment to the foreclosurefrequencies. We applied adjustment factors of 2.25x, 2.00x, and 1.75x to the foreclosurefrequencies for loans using 1-11 months, 12-23 months, or 24 months or more of incomeverification, respectively.

- About 17.6% of the loans are backed by property types other than single-family homes andPUDs, such as condominiums, co-operatives, two-to-four family houses, and other propertytypes. We applied an adjustment to their foreclosure frequencies to account for this risk.

- The loan purpose for 21.5% of the loans by balance is cash-out refinance.

- Non-QM loans, which have an increased risk of ATR challenges and associated loan losses,make up 61.6% of the pool. We applied an adjustment to loss severities to account for this risk,which increased our loss coverage estimates for each rating category.

- A portion of the mortgage loans (23.2% of the pool balance) were made to borrowers with FICOscores below 700 (including our assumption of a 686 FICO score for borrowers who lack a FICOscore). The mortgage pool's loss estimate was increased to account for the increased defaultrisk of these loans.

- The respective originators make the R&Ws (or, with respect to the Acra mortgage loans, R&Wsrestated by the sponsor that were originally made by Acra Lending) with the sponsor,backstopping the originators' repurchases in the event of an insolvency of the originator, whichis a strength. However, the early payment default covenant is not applicable to loans originatedby Acra Lending, which we consider as a weakness. Further, the R&W framework is somewhatweak because the controlling holder (the majority owner of the class X certificates and, initially,an affiliate of the sponsor) is not required to test for breaches on all loans that go delinquent orsustain a realized loss, although it has the option. However, the controlling holder is obliged toreview all ATR-notice (a violation of the ATR rules raised) loans that suffer a realized loss, and itmay have conflicting interests vis-à-vis the certificateholders. The risks and concerns with theframework are somewhat mitigated by third parties performing 100% due diligence on theloans and the sponsor (through a majority-owned affiliate of the sponsor) holding the first-losspiece and retaining risk via a material net economic interest of not less than 4.0% of the capitalstructure. Consequently, we applied an R&W adjustment that increased our loss expectationsat all rating categories by a 1.10x factor.



2668248


Credit Analysis And Assumptions

Our analysis of the GCAT 2021-NQM3 collateral pool considered a number of factors, includingcertain loan-level characteristics detailed below.

Documentation type

The sponsor guidelines allow income verification using paystubs, W-2s/W-2 equivalents, taxreturns, written verification of employment, CPA-prepared P&L statements, and personal orbusiness bank statements. The sponsor also offers asset depletion programs that considerborrower accumulated assets in calculating income to qualify the borrower. Cash flow investorloans are underwritten, in part, to a DSCR, which is calculated using actual or estimated rent.

Table 5

Documentation Type (Income Verification Type/Length)

Loancount

(no.)

Currentbalance

(%)

Alternative IncomeVerification Length

(WA # of months)

Foreclosurefrequency

adjustmentfactors (x)

AAA' Foreclosurefrequency without

pool adjustmentfactors (%)

Full Documentation

Appendix Q/qualifiedmortgage

1 0.03 19.40

Full (24+ months) 53 15.41 1.00 20.27

Full (24+ months) WVOE only 20 2.30 1.00 23.45

Full (12-23 months) 2 0.53 1.25 31.79

Full (1-11 months) 1 0.09 1.50 13.90

Alternative documentation(i)

24+ months

Business bank statements 41 10.74 24.0 1.75 30.66

Personal bank statements 9 2.17 24.1 1.75 37.01

P&L statements 2 0.22 24.0 1.75 63.33

12-23 months



P&L statements 39 5.77 12.0 2.00 43.66

1-11 months



P&L statements

Other documentation

Other (DSCR) 128 16.21 3.15-6.00 68.37

Other (applied 0.00 DSCR) 5 0.66 6.00 52.51



2668248


Table 5

Documentation Type (Income Verification Type/Length) (cont.)

Loancount

(no.)

Currentbalance

(%)

Alternative IncomeVerification Length

(WA # of months)

Foreclosurefrequency

adjustmentfactors (x)

AAA' Foreclosurefrequency without

pool adjustmentfactors (%)

Other (assetunderwriting/depletion)

12 3.18 3.00 28.91

(i)The documentation source may include other secondary documentation types. WA--Weighted average.

For 76 loans (approximately 18.3% of the pool balance), traditional (full) documentation was usedfor fully verifying and calculating the borrowers' qualifying income (e.g., written verification ofemployment, pay stubs, W-2s, personal and business tax returns, and IRS transcripts). We applieda documentation type adjustment factor ranging from 1.00x to 1.50x, depending on the length ofthe income verification, for these loans.

We classified all loans to borrowers that used nontraditional sources of income documentation,such as bank statements (business or personal) or CPA-prepared P&L statements documentingincome as alternative documentation loans. Alternative documentation was used on 294mortgage loans (61.6% of the pool balance), with most borrowers using 12-23 months of incomeverification. We view income verification using alternative documentation to be a weaker standardthan full documentation of income. Consequently, we increased our loss coverages for these loansby an adjustment factor ranging from 1.75x to 2.25x.

One hundred and twenty eight loans in the pool (16.2% by pool balance) were underwritten to alending program that considers investment property cash flow. We classified these loans as"other" documentation loans with a DSCR flag and applied a 3.15x-6.00x adjustment factor to theforeclosure frequencies based on the provided DSCR calculation. The DSCR calculations providedby the issuer for these loans ranged from 0.45 to 2.67.

We classified five investment property business purpose loans (0.66% by balance) that wasprimarily underwritten to FICO score and LTV ratio (rather than DSCR) as "other" documentationtype. We evaluated this non-DSCR loan as if it were a DSCR loan with a DSCR of zero; our lossmodel applied a 6.00x adjustment factor to the foreclosure frequencies for this non-DSCR loan.

Twelve loans in the pool (3.2% of the pool balance) were underwritten, in full or in part, by alending program that considers accumulated assets rather than a verified income stream. Weclassified these loans as other loans and applied a 3.00x adjustment to the foreclosurefrequencies.

Of the 89 Change Lending- and Quontic-originated loans (which are ATR exempt because they areCDFIs), we classified 34 as full, 49 as alternative, and the remaining six as other documentationtype loans. Both Change Lending and Quontic guidelines generally specify using a writtenverification of employment or the more commonly used W-2/tax statements to document a wageearner's income. As such, we treated these loans as "full" documentation type loans. Forself-employed borrowers, the Quontic guidelines generally specify using a CPA-prepared12-month P&L statement for income documentation, while Change Lending's guidelines specify12 months of bank statements (business or personal); we treated these loans as "alternative"documentation type loans.



2668248


Prior credit event (PCE) classification and analysis

We focused primarily on prior bankruptcy, foreclosure, short sale, and deed-in-lieu events (24months from the cut-off date for bankruptcy discharges or dismissals and 36 months from thecut-off date for housing-related events). For loans to borrowers with more seasoned PCEs, webelieve the risks associated with those PCEs are appropriately reflected in the updated FICO. Fourloans (0.5% of the pool balance) fell into the PCE category, as defined in our criteria. We applied a1.01x pool-level PCE related loss coverage adjustment factor.

QM and ATR standards

The CFPB issued final regulations for mortgage loans with applications submitted on or after Jan.10, 2014, specifying the standards for a QM. Based on the designation provided by the sponsor,61.6% of the loans (by balance) are categorized as non-QM/compliant (see table 6).

Table 6

Qualified Mortgage Breakout

QM status Pool balance ($) % by pool balance Loan count (no.) Weighted average FICO score

QM/non-HPML -- -- -- --

QM/HPML 85,733 0.03 1 740

Non-QM/compliant 178,947,406 61.59 281 733

Not covered/exempt 111,522,705 38.38 234 736

QM--Qualified mortgage. HPML--Higher-priced mortgage loan.

Under the ATR rule, the originator and any assignee are jointly and separately liable for certaindamages that may be incurred from noncompliance with the rule (see Appendix I of "MethodologyAnd Assumptions For Rating U.S. RMBS Issued 2009 And Later," published Feb. 22, 2018). Foreach loan subject to the rule, we applied an adjustment to loss severities to account for this risk.The data the issuer provided to S&P Global Ratings, including additional fields that validate theloan's QM designation, were reviewed by the due diligence firms under the third-party duediligence firms' scope to verify that documentation exists to support the QM designation. Also, inconjunction with our aggregator review, we concluded that BRM II's processes address the ATRrisks.

Servicer advancing obligations

For any loan that is not in forbearance and serviced by the servicer, NewRez LLC (doing businessas Shellpoint), Shellpoint must advance delinquent P&I on any delinquent mortgage loan until it iseither greater than 90 days' delinquent (limited P&I advancing) or the P&I advance is deemedunrecoverable by the related servicer. In the event that the servicer fails to advance P&I, themaster servicer, Nationstar, is obligated to make those advances unless the master servicerdetermines that the P&I advance is nonrecoverable. To the extent that the servicer and the masterservicer fail to make a required P&I advance, the securities administrator will be obligated to makethat required P&I advance.

Approximately 43.5% of the loans by pool balance as of the cut-off date are serviced by Citadel.Neither Citadel nor any other transaction party is required to advance any delinquent interest or



2668248


principal on these mortgage loans.

Unlike P&I advances, the servicers must make advances of delinquent taxes and insurance (andother property preservation advances) on any delinquent mortgage loan until the related propertyis liquidated or the related servicer deems the advance to be unrecoverable. We incorporated thelimited P&I advancing into our loss severity calculations.

Borrowers with multiple loans

We did not make an additional adjustment to the loss coverage and the tail risk analysis due toborrowers with multiple loans in the transaction because only six borrowers have multiple loans.These loans combined represent 1.9% of the pool balance.

Structural Features

The GCAT 2021-NQM3 structure has a mix of pro rata and sequential features. Principal is paid prorata among the senior classes (subject to passing cumulative loss and delinquency trigger tests)and then sequentially to the mezzanine and subordinate classes. In the periods when thecumulative loss or delinquency trigger test fails, principal is first used to pay any unpaid interestand interest carryforward amounts (to the extent not paid after allocation of the interestremittance amount) sequentially first concurrently to class A-1 and class A-1X) and then class A-2certificates, then pay principal sequentially to classes A-1 and A-2; and then pay any unpaidinterest and interest carryforward amounts to the classes, followed by principal to that class untilreduced to zero--sequentially to the A-3, M-1, B-1, B-2, and B-3 certificates in that order, withboth interest and principal paid to a class before payments to the next class.

Since the class A-1, A-2, and A-3 certificates can receive principal pro rata, the amount ofprotection to the class A-1 and A-2 certificates can decline over time. In our analysis, thedelinquency or cumulative loss trigger may help protect the more-senior classes by allowing thepayment mechanism to switch to sequential earlier, thus preserving subordination and requiringless upfront credit enhancement.

The transaction also uses monthly excess cash flow to cover the current period's realized lossesand to reimburse any previously applied realized loss amounts.

The securities administrator will make monthly distributions of interest from the interestremittances and principal from principal remittances (see tables 7-9).

Table 7

Interest Payment Waterfall

Priority Payment

1 Interest and interest carryforward amounts(i) first concurrently to class A-1 and A-1X certificates, andthen sequentially to the class A-2, A-3, M-1, B-1, B-2, and B-3 certificates.

2 Any remaining amount paid as part of monthly excess cash flow.

(i)Interest carryforward amount are deferred interest payments that accrue interest at the lower of the respective coupon and the net WAC rate.Our ratings address the expectation for full payment of all interest and interest carryforward amounts by the final maturity date.WAC--Weighted average coupon.



2668248


Table 8

Principal Payment Waterfall

Priority Payment

If the delinquency and cumulative loss trigger tests pass

1 Unpaid interest and interest carryforward amounts first concurrently to class A-1 and class A-1Xcertificates, and then sequentially to the class A-2, and A-3 certificates.

2 Principal pro rata to the class A-1, A-2, and A-3 certificates until reduced to zero.

3 Unpaid interest and interest carryforward amounts to the class M-1 certificates .

4 Principal to the class M-1 certificates until reduced to zero.

5 Unpaid interest and interest carryforward amounts to the class B-1 certificates.

6 Principal to the class B-1 certificates until reduced to zero.






If the delinquency or cumulative loss trigger tests fail

1 Unpaid interest and interest carryforward amounts concurrently to the class A-1 and class A-1Xcertificates.

2 Unpaid interest and interest carryforward amounts to the class A-2 certificates.

2 Principal to the class A-1 certificates until reduced to zero.


5 Unpaid interest and interest carryforward amounts to the class A-3 certificates.


7 Unpaid interest and interest carryforward amounts to the class M-1 certificates.

8 Principal to the class M-1 certificates until reduced to zero.








Table 9

Monthly Excess Cash Flow Waterfall

Priority Payment

1 Sequentially to the class A-1, A-2, A-3, M-1, B-1, B-2, and B-3 certificates up to the realized loss amountfor the current period until their respective class certificate balances are reduced to zero.

2 Sequentially, up to the cumulative applied realized loss amount to the class A-1, A-2, A-3, M-1, B-1, B-2,and B-3 certificates until their respective certificate balances are reduced to zero.



2668248


Table 9

Monthly Excess Cash Flow Waterfall (cont.)

Priority Payment

3 To the cap carryover reserve account, from monthly excess amount otherwise distributable to the class Xcertificates; the aggregate cap carryover amount(i) for classes A-1, A-1X, A-2, A-3, M-1, B-1, and B-2 forsuch distribution date; and, from the cap carryover reserve account, any unpaid cap carryover amountsfirst concurrently to class A-1 and class A-1X certificates, and then sequentially to the class A-2, A-3,M-1, B-1, and B-2 certificates.

4 Sequentially, any amounts due to the class X certificates, then reimburse ETE amounts that exceedannual caps, pro rata, to the relevant transaction parties, then any remaining amounts to the interestreserve account, and then any remaining amounts to the class R certificates.

(i)The cap carryover amount is the positive difference between the interest that would have accrued at the coupon rate (without regard to thepool's net WAC rate) and what was actually due based upon the net WAC rate. Any prior unpaid cap carryover amounts also accrue at the couponrate without regard to the net WAC rate. Our preliminary ratings do not address the payment of cap carryover amounts. PSA--Pooling andservicing agreement. WAC--Weighted average coupon. ETE--Extraordinary trust expenses.

The interest remittance amount includes the interest collected from, or advanced on behalf of,borrowers (including interest payments that accompany prepayments, any compensating interestand interest portions of liquidation proceeds [minus expenses], subsequent recoveries,termination prices, redemption prices, and repurchase amounts) minus servicing, masterservicing, trustee, loan-data agent, and custodial fees, as well as the servicer advancereimbursements permitted under the pooling and servicing agreement, reimbursable expensesincurred by the controlling holder, and extraordinary expenses, which are generally capped at$300,000 annually. Although the extraordinary expenses are passed through as reducedcontractual interest due to certificateholders, we ran these expenses at a certain percentage oftheir capped amounts (as described in the Interest Stresses section below). We also consideredthe extraordinary expenses when analyzing projected interest reduction amounts (as described inthe Imputed Promises Analysis section below).

Principal remittance amounts include the principal collected from or advanced on behalf ofborrowers (including prepayments, principal portions of liquidation proceeds [minus expenses],subsequent recoveries, termination price, redemption price, and repurchase amounts) minusfees, including extraordinary trust expenses that could not be paid from interest collections.

The interest remittance amounts, and principal remittance amounts will not include anypre-closing deferred amounts collected by the servicers.

Interest on the class A-1, A-1X, A-2, A-3, M-1, B-1, and B-2 certificates is based on the lower ofthe coupon rate on the certificates and the net weighted average coupon (WAC) rate. Further, classA-1X, which will be outstanding for 36 months, will have a notional amount equal to the lesser of(a) the balance of the A-1 certificates and (b) the notional amount set forth on a schedule for therelated accrual period. Interest on class B-3 is the net WAC rate. The net WAC rate is defined asthe weighted average of the mortgage interest rates of the loans, net of fees and extraordinaryexpenses, weighted based on the loans' stated principal balances. Our preliminary ratingsaddress the lower of these two rates.

Under the transaction documents, the issuer can defer interest payments on these securities. Afailure to pay the interest amounts due on the securities will result in the interest being deferred.Deferred interest (the interest carryforward amount) accrues interest at the lower of the fixed rateand net WAC rate for each class. Our preliminary ratings address the expectation for P&Ipayments (including interest carryforward amounts) by the certificates' final maturity date.

However, our preliminary ratings do not address the expectation for payment of the cap carryover



2668248


amounts (i.e., the difference between the coupon and the net WAC where the coupon exceeds thenet WAC). These amounts are subordinated in the payment priority. In our view, neither the initialcoupons on the certificates nor the initial net WAC rates are de minimis, and the nonpayment ofthe cap carryover amount is not considered an event of default under the transaction documents.Therefore, in line with our criteria for imputed promises, we do not consider whether these capcarryover amounts are paid in our cash flow analysis.

In GCAT 2021-NQM3, the mezzanine and subordinate certificates are paid principal sequentiallyafter all senior certificates have been paid. Unlike the credit enhancement seen inshifting-interest RMBS structures, which may deplete due to scheduled and prepaid principalpaid to the subordinate classes, the credit enhancement in GCAT 2021-NQM3 is not depletedbecause principal payments are not made to these classes unless they are the most senior classoutstanding.

Although principal is paid pro rata among the senior classes from the start and there is no definedcredit enhancement floor that would switch the senior classes' payment priority to sequential, weare comfortable that the transaction is adequately enhanced for the assigned preliminary ratings,taking into account any tail-risk considerations given that the transaction starts with 12.4%enhancement for the senior classes, which then grows as a percentage of the current balance asthey get paid down (see the Large Loan And Tail-Risk Considerations section). Additionally, thedelinquency trigger and cumulative loss rate trigger (see tables 10 and 11) protect the more seniorclasses in tail-risk situations if defaults were to increase much later in the transaction's life (aback-ended default curve) by switching the payment priority among the senior classes tosequential.

Table 10

Cumulative Loss Trigger Event

Distribution date occurring in the followingperiods

Applied realized loss amounts since the closing date (as a % of thecut-off date pool balance)

July 2021 through June 2024 2.00



July 2026 and thereafter 7.00

Table 11

Delinquency Trigger Event

Distribution date occurring in thefollowing periods

Six-month average of 60+ day delinq. plus loans modified in past 12 monthsincluding any deferrals (as a % of the current pool balance)



July 2026 and thereafter 30.00

Delinq.--Delinquent.

If the aggregate class balance of the certificates exceeds the pool balance, the resulting excess(the applied realized loss amount) is applied in sequential order to the class B-3, B-2, B-1, M-1,A-3, A-2, and A-1 certificates until each class' principal balance has been reduced to zero.

If the pool balance exceeds the aggregate class balance of the certificates (after the allocation of



2668248


principal payments and monthly excess cash flow to pay down the certificates), the balances ofthe class A-1, A-2, A-3, M-1, B-1, B-2, and B-3 certificates will be written up sequentially in thatorder to the aggregate amount of applied realized losses previously allocated.

Geographic Concentration

S&P Global Ratings analyzes the pool's geographic concentration risk based on theconcentrations of loans in each of the CBSAs defined by the U.S. Office of Management andBudget (see Appendix II of "Methodology And Assumptions For Rating U.S. RMBS Issued 2009 AndLater," published Feb. 22, 2018). In this transaction, the top five CBSAs account for 36.4% of theaggregate pool. We applied a geographic concentration adjustment factor of 1.02x to our base losscoverage estimate (see table 12).

Table 12

Geographic Concentration

CBSA code(i) CBSA State % by balance

31084 Los Angeles-Long Beach-Glendale (Metropolitan Division) California 12.39

33124 Miami-Miami Beach-Kendall (Metropolitan Division Florida 7.62

11244 Anaheim-Santa Ana-Irvine (Metropolitan Division California 6.11

22744 Fort Lauderdale-Pompano Beach-Sunrise (MetropolitanDivision)

Florida 5.50

35614 New York-Jersey City-White Plains (Metropolitan Division) New York 4.76

Top five -- -- 36.39

(i)CBSA code refers to the metropolitan division code, if available. CBSA--Core-based statistical area (includes metropolitan statistical areasand metropolitan divisions where defined, as well as micropolitan statistical areas).

Large Loans And Tail-Risk Considerations

As the number of loans in the transaction decreases, the effect of a single loan's losses becomesgreater. If conditional prepayment rates are slow and collateral pool losses are not realized untillater in a transaction's life (back-loaded losses), pro rata pay mechanisms can then leave thesenior certificates exposed to event risk later in the transaction's life (for more information on tailrisk in RMBS transactions, see "Older RMBS Transactions Face Increased Tail Risk As Their PoolsShrink," published Aug. 9, 2012).

To mitigate this risk, certain transactions provide for a credit enhancement floor, specifyingprincipal payments not be made to mezzanine or subordinate classes if the credit supportavailable to the senior classes falls below a threshold. GCAT 2021-NQM3 does not explicitlyprovide a credit enhancement floor. However, due to the sequential payment mechanism to themezzanine and subordinate classes, which make up 12.4% of the capital structure, thepreliminary 'AAA (sf)', 'AA (sf)', and 'A+ (sf)' rated classes effectively have a floor of 12.4% initially.Since subordination to the senior classes is locked out from receiving any principal paymentsreceived on the mortgage loans, the 12.4% should be available to absorb losses in the eventdefaults begin to occur after an extended period of benign performance, which is the scenario ourtail risk analysis is intended to address. Further, when cumulative losses or delinquencies trip thecumulative loss or the delinquency triggers, the payment priority becomes fully sequential.

To analyze the appropriateness of this effective credit enhancement floor, we use an approach



2668248


outlined in "Methodology And Assumptions For Rating U.S. RMBS Issued 2009 And Later,"published Feb. 22, 2018. Per this approach, instead of focusing on the largest loans by balance atissuance, we risk-weight the loans in the transaction by focusing on those loans with the largestexpected loss exposure, assuming default. The resulting projected losses at the 'AAA (sf)', 'AA (sf)'and 'A+ (sf)' level are below the 12.4% effective floor provided in the deal.

After considering the credit enhancement provided in the transaction, the cumulative loss anddelinquency trigger definitions and the certificates' expected paydown, we believe the rated seniorcertificates are sufficiently protected from tail risk as the transaction seasons.

Mortgage Operational Assessment (MOA)

Angelo, Gordon and Co. L.P. (Angelo Gordon) was founded in 1988 and the structured creditplatform at Angelo Gordon began operations in 2008. Blue River Mortgage TRS (BRM) was formedin 2017 as a privately held real estate investment trust that is managed by Angelo Gordon andowned by certain funds managed by Angelo Gordon. Subsequently in July 2020, BRM II wasformed as an additional vehicle to acquire non-QM mortgage loans via the same policies,procedures, controls, and staffing of BRM. For the purposes of this transaction we incorporatedour MOA of BRM (detailed below) to BRM II.

Given its management and ownership structure, BRM benefits from various resources provided byAngelo Gordon, such as employees, technology, oversight, and the general expertise of a firm withover 30 years of experience and more than $43 billion in assets under management (AUM). BRMacquires non-QM mortgages through bulk and flow purchases with the mortgage servicing rightstypically released. NewRez LLC (doing business as Shellpoint), an S&P Global Ratings selectservicer; AmWest Funding Corp.; Metro City Bank; Royal Business Bank; and Citadel ServicingCorp., currently service the loans in the BRM portfolio.

Since December 2017, BRM has purchased non-QM loan pools from 10 sellers. It may onboardadditional sellers to the platform that meet its targeted profile of proven, larger originators with asufficient operating history. Prospective sellers are reviewed via a multistage process and must beapproved unanimously by the operational committee, which consists of three teams (pricing andportfolio management, transaction management and operations, and credit and legal) prior tobeing submitted to the management committee, which comprises senior members of AngeloGordon's structured credit team, for final approval. In addition to this approval, BRM engages athird-party firm to conduct an operational review of sellers with significant volume. Although BRMhas a short history of purchasing non-QM loans, the parent company and the management teamhave extensive experience purchasing re-performing and non-performing loans, in addition toother real estate assets.

Our overall ranking reflects our qualitative and quantitative review of BRM. Our qualitative reviewis based on our assessment of three primary focus areas covered during the operational review:

- Management and organization, which includes risk management and financial position;

- Loan purchase and aggregation, which includes property valuation processes; and

- Internal controls, which encompasses operational reviews of originators, pre-purchase dataquality, post-purchase quality control, and regulatory compliance.

For our quantitative analysis, we reviewed acquisition volume, loan characteristics, and loanperformance history, including delinquencies, early payment defaults (EPDs), and repurchases.

Our AVERAGE qualitative subranking and LIMITED quantitative subranking reflects our



2668248


assessment of the following strengths and weaknesses.

Strengths:

- The management team's averaged industry experience of over 20 years.

- BRM's relationship with and support from Angelo Gordon, an established investor in credit andreal estate with over $43 billion in assets under management across a broad range of strategiesand experience in reperforming and nonperforming loans.

- Due diligence is conducted on 100% of loans and includes a full review of credit, compliance,property valuation, and fraud.

- The robust internal controls; critical risk decisions require unanimous approval of theoperational committee members or subsequent approval by the management committee.

Weaknesses:

- Loan performance hasn't experienced an economic downturn and therefore is untested.

- The post-purchase review process is limited, though this is mitigated via a pre-purchasethird-party due diligence review of all loans.

Mortgage originator concentration

In addition to an MOA of BRM (and its application to BRM II for purposes of this transaction), wereviewed historical performance data for loans from originators with exposures above 20% in theGCAT 2021-NQM3 pool. Specifically, loans purchased from Acra Lending and Arc Home LLC, whichrepresent approximately 43.5% and 31.5% of the pool balance, respectively.

Based on the results of this analysis, we concluded that the historical performance of these loanswas comparable to the performance of the loans in the BRM and BRM II total portfolio.

We applied a 1.05x adjustment to the loss coverage at all rating categories to reflect our view ofBRM II's residential mortgage acquisition platform.

Third-Party Due Diligence Review

The third-party due diligence providers--Clayton Services LLC, Consolidated Analytics LLC, DigitalRisk LLC, Evolve Mortgage Services LLC, and Recovco Mortgage Management LLC (which are all onour list of reviewed providers)--performed due diligence on 100% of the pool's loans. Their reviewsencompassed credit (underwriting) compliance, property valuations, regulatory compliance(where applicable), and data quality.

Some loans fell within the scope of the TILA-RESPA Integrated Disclosure (TRID) rule. For theseloans, the third-party firms followed the Structured Finance Assn. (SFA) RMBS 3.0 TRIDCompliance Review Scope in conducting their final loan reviews (see "Standard & Poor'sComfortable With SFIG Draft Proposal Regarding TRID Due Diligence," published April 25, 2016). Inaccordance with our criteria, we adjust our loss expectations based on our view of the firms'findings (see Appendix III of "Methodology And Assumptions For Rating U.S. RMBS Issued 2009And Later," published Feb. 22, 2018).

Highlights of some of the findings include the following:



2668248


Compliance with underwriting guidelines (credit review)

- Most of the due diligence firms' initial findings were resolved by the receipt of missinginformation in the trailing documents.

- All loans in the pool received a credit risk grade of 'A' or 'B'.

Property valuation review

- All but one loan received a final property valuation review risk grade of 'A' or 'B'. One loanreceived a property valuation grade C because the value could not be supported within 10% ofthe original appraisal amount

Regulatory compliance review

- Most of the due diligence firms' initial findings were resolved by the receipt of missinginformation in the trailing documents.

- Many exceptions resulted from missing documents to complete the ATR/QM or TRID reviews.

Data quality review

- A final tape was provided with updated/corrected data, and no outstanding data issues werenoted.

After reviewing the third-party due diligence results, we applied a neutral adjustment of 1.00x tothe loss coverage at all rating categories.

R&Ws

Our review of the R&Ws for GCAT 2021-NQM3 focused on whether the representations made bythe R&W providers (the originators or, in the case of Acra Lending mortgage loans, the sponsor)who restates all the R&Ws as of the closing date originally made by Acra Lending in the relatedunderlying mortgage loan purchase agreement, were substantially consistent with the set ofrepresentations we published as part of our criteria (see Appendix IV of "Methodology AndAssumptions For Rating U.S. RMBS Issued 2009 And Later," published Feb. 22, 2018). Weevaluated the strength of these R&Ws and considered whether any breach could have a materiallyadverse impact for the certificateholders. If the R&Ws in the transaction documents do notaddress the issues in our published R&W framework, we will determine whether we believe it isappropriate to assess additional credit enhancement. Lastly, we will consider the R&W providers'ability to fulfill their obligations in the event of a breach.

The collateral pool consists of loans from five originators. In this transaction, the originators(except for Acra Lending) make the R&Ws with the sponsor backstopping the originator in theevent of the originator's insolvency. For Acra Lending mortgage loans, the sponsor restates all therepresentations and warranties originally made by Acra Lending as of the closing date. Wereviewed the representations made by the representation providers that were assigned to thetrust or restated by the sponsor for the benefit of the certificateholders. The originators, like the



2668248


sponsor, are unrated entities that may be financially unable to repurchase loans if the need arises.We consider the R&W framework to be weaker than the R&W framework seen in prime jumbotransactions because the review of the breach is not automatic, but is instead, at the option of thecontrolling holder (initially, the affiliate of the sponsor)--unless a violation of the ATR rules israised on a loan that suffers a realized loss. In this case, the controlling holder has to review theloan for a breach. However, the R&Ws are in line with other non-QM RMBS transactions.

The R&Ws generally are consistent with our criteria. Knowledge qualifications are limited and donot significantly alter the R&W or remedy. The R&Ws are generally made as of the closing dateand, where applicable, the sponsor provides representations during any gap period. The R&W willremain in effect for the transaction's life. However, given that the statute of limitations for R&Wclaims under New York law is generally six years from the date a representation is made, therecould effectively be an expiration date on the R&Ws. The early payment default covenant for allloans (excluding those originated by Acra) is in line with what we typically look for in a ratedtransaction, which generally calls for the R&W provider to repurchase a mortgage loan when theborrower fails to make any of the first three monthly payments due after the origination dateunless the delinquency resulted from a servicing issue that has subsequently been, or will be,corrected. In GCAT 2021-NQM2, if a borrower fails to make any of the first three monthly payments(not including any payment subject to a forbearance due to the COVID-19 pandemic) by the lastday of the month it is due (or in some cases fails to make the first payment by the last day of themonth it is due), measured generally from BRM's purchase date, the applicable originator isrequired to repurchase the loan. The early payment default covenant is not applicable to loansoriginated by Acra Lending. However, the sponsor may, with respect to mortgage loans originatedby Acra Lending, purchase such mortgage loans as a result of an early payment default if AcraLending agrees to purchase such mortgage loan from the sponsor. Early payment default relatedrepurchases are not backstopped. We took this into account in our overall assessment of the R&Wand the related adjustment factor applied.

The applicable originator and/or sponsor must appropriately remedy any ensuing R&W breach if ithas a materially adverse impact on the loan by curing the breach, or repurchasing the mortgageloan or REO property, as applicable, at the purchase price. The enforcement mechanism for R&Wbreaches includes provisions for a breach review at the option of the controlling holder--either byan independent reviewer or by the controlling holder itself--for any loan that experiences arealized loss. However, loans with realized losses for which ATR notices have been received areautomatically reviewed. If a loan is judicially determined to have a TRID finding, or if the custodiandetermines a loan has defective or missing documentation, it must be cured or repurchased.

Dispute resolutions are ultimately subject to arbitration proceedings, if necessary, in case theoriginator and/or the sponsor disputes the controlling holder's determination. The costs of thearbitration will be paid by the controlling holder, subject both to reimbursement for thoseexpenses from the assets of the issuing entity and the annual cap, to the extent the controllingholder wins the arbitration.

The transaction also has a provision in which 25.0% of the certificateholders (the directingcertificateholders) can start a review on a loan with a realized loss in case the controlling holdereither does not review the loan for a breach or determines that the related originator is notobligated to cure/repurchase the loan. Any costs for that review will be borne by these directingcertificateholders and may be recovered from the originator/sponsor if, eventually, the directingcertificateholders prevail following any arbitration. However, if the originators do not reimbursethe costs incurred by the directing certificateholders, those costs may be reimbursed from thetrust.

Although the MOA's result reflects a solid aggregation platform, in our view, parties with



2668248


potentially limited repurchasing ability are providing R&Ws. Therefore, in conjunction with theweaknesses in the framework as identified above, we applied a 1.10x loss coverage adjustment atall rating levels to compensate for these risks. We believe this adjustment is appropriate in thecontext of the 100% due diligence performed on the loans and the collateral's relative creditquality.

Cash Flow And Scenario Analysis

We reviewed the transaction structure and performed a cash flow analysis to simulate variousrating stress scenarios to determine the preliminary ratings for each class consistent with ourcriteria, accounting for the available credit enhancement (see tables 13 and 14). We analyzed avariety of scenarios for each rating category, including combinations of the following:

- Front- and back-loaded default timing curves;

- Two-year recovery lag assumptions;

- Fast and slow prepayment assumptions;

- Extraordinary trust expense stresses;

- High, low, and forward/flat interest rate curve assumptions; and

- Delinquency assumptions to stress liquidity for lack of advancing on Citadel serviced loans andpotential forbearance on the total pool.

Table 13

Cash Flow Assumptions

Scenario

AAA AA A+ BBB BB B

Recovery lag (mos.) 24 24 24 24 24 24

Servicer advancing No advancing on loans during the recovery lag period.

Prepayments (%)(i)

Low CPR 1 2 3 4 5 6

High CPR 20 20 20 20 20 20

Scenario 1: Delinquencycurve

Standard delinquency curve for testing triggerswithout cash flow stress.


Delinquencies at 35% for first six months to stressliquidity and triggers followed by standarddelinquency curve to test triggers.


20% of the loans assumed to default under thedefault timing curve in a given month and remaindelinquent for three months in addition to expecteddefaults.

Extraordinary trustexpenses (% of cappedamount)(ii)

100.00 100.00 95.00 40.00 30.00

Foreclosure frequency(%)

42.53 37.52 32.67 23.47 16.72 10.15

Loss severity (%) 48.79 43.84 34.74 29.40 25.72 23.15



2668248


Table 13

Cash Flow Assumptions (cont.)

Scenario

AAA AA A+ BBB BB B

Loss coverage (%) 20.75 16.45 11.35 6.90 4.30 2.35

(i)Using a standard prepayment convention. CPR--Conditional prepayment rate. N/A--Not applicable.

Notwithstanding the use of excess interest within the transaction structure, we applied front- andback-loaded rather than bulleted (e.g., a semiannual or an annual lump sum) default timingcurves in our analysis. This reflects our view of the potential volatility of cash flows, given the loansare newly originated by a reviewed aggregator, subject to third-party due diligence and includestructural considerations, such as pro rata principal allocations among classes A-1, A-2, and A-3,and partial P&I advancing by the servicer.

We applied the foreclosure frequencies, loss severities, and combinations of the stresses notedabove in our cash flow runs, and we observed some periodically missed interest due to theliquidity stress associated with no advancing. To pass our applicable rating-specific stresses, theinterest carryforward amounts resulting from any missed interest payments on the securitieshave to be paid in full by the maturity date. All carryforward interest was paid back with interestunder the applicable rating-specific stresses in our cash flow projections. The results show thateach preliminary rated class in the transaction is enhanced to a degree consistent with theassigned preliminary ratings.

Table 14

Structural Assessment

ClassPreliminaryrating

Initial classsize (%)

Initial creditenhancement (%)

Loss coverage(%)

Percentage point differencebetween credit enhancement and

loss coverage

A-1 AAA (sf) 73.45 26.55 20.75 5.80

A-1X(i) AAA (sf) 73.45 N/A N/A N/A

A-2 AA (sf) 7.35 19.20 16.45 2.75

A-3 A+ (sf) 6.80 12.40 11.35 1.05

M-1 BBB (sf) 6.70 5.70 6.90 (1.20)

B-1 BB (sf) 2.55 3.15 4.3 (1.15)

B-2 B (sf) 1.80 1.35 2.35 (1.00)

B-3 NR 1.35 0.00 N/A N/A

(i)Class A-1X will have a notional amount equal to the lesser of (a) the balance of class A-1 immediately prior to such distribution date and (b)the notional amount set forth on a schedule for the related accrual period. After the 36th distribution date, the notional amount of the A-1Xcertificates will be zero. NR--Not rated. N/A--Not applicable.

Servicer stop advance stresses

Although the transaction documents provide for up to three months of P&I advance obligation onapportion of the pool, which would address temporary delinquencies, we assumed that no P&Iadvances were being made in our cash flow projections on loans projected to default that have notyet been liquidated (we assume a 24-month lag between default and liquidation). In addition, to



2668248


stress temporary delinquencies for no P&I advancing on loans serviced by Citadel, based on ourcriteria, we assumed temporary delinquencies equal to 20.00% of the loans assumed to default(under our default timing curve in a given month) would remain delinquent for three months beforecuring. Our cash flow projections take into account these liquidity stresses and the transaction'sability to make monthly interest payments and, if necessary, interest carryforward amounts by thefinal maturity date on the preliminary rated classes. We also modeled a delinquency curve, basedon our criteria, for the purpose of testing the delinquency trigger.

To address the potential liquidity stress to cash flows due to loans entering forbearance in light ofthe current COVID-19 pandemic crisis for which the P&I advancing party is not obligated toadvance monthly P&I payments, we applied an additional delinquency stress scenario. Weassumed 35.00% of the closing pool balance to be delinquent for the first six months with any P&Ipayments related to this delinquent portion coming back to the transaction after all defaults havebeen passed through to the transaction (approximately 144 months).

WAC deterioration stress

To address the potential for a pool's WAC to decline over time as higher coupon loans prepay ordefault, we stress the pool's projected cash flows by reducing the interest accrued on the assets.Where appropriate, we review the distribution of loan coupons in the pool, based on measuressuch as the standard deviation, interquartile range, and maximum/minimum ranges to assess thepool's homogeneity with respect to loan coupons.

Generally, the stress is based on the pool's WAC at the time of analysis versus 10 years later,based on an assumed reduction in the pool balance of 10.00% per year applied to the loans withthe highest coupons. This WAC difference is the maximum WAC deterioration assumed for thepool. The stress applied starts at zero in the transaction's first month and increases linearly eachmonth to the maximum through year 10, at which point, it remains constant at the maximumthrough the deal's remaining life. This stress is applied in all cash flow stress scenarios at allrating levels. For this mortgage pool, we applied a maximum WAC deterioration of 0.95%.

Interest stresses

All of the rated certificates have coupons subject to the net WAC-rate cap, as is the case for mostpost-2009 transactions that we have rated. If the net WAC rate decreases below the cap, theinterest due to the certificates will decrease by a similar amount.

In this transaction, extraordinary trust expense payments reduce the net WAC rate, whicheffectively allocates the extraordinary trust expenses pro rata across all senior and subordinatecertificateholders by reducing their interest payments by the amount of the extraordinary trustexpenses paid (subject to the annual cap). Although the extraordinary trust expenses are passedthrough as reduced contractual interest due to certificateholders, we ran these expenses fromperiod 13 to 60 (four years) at a certain percentage of the capped amounts as specified by ourcriteria to test any impact on the securities due to their dependence on excess spread as a form ofcredit enhancement and the presence of certain structural features, such as limited P&Iadvancing. We also took this approach because interest payments on the securities aredeferrable.

Loans with interest rates indexed to six-month and one-year SOFR represent approximately 4.1%of the pool balance. We applied flat stresses to the SOFR rates in accordance with our criteria (see"Methodology To Derive Stressed Interest Rates In Structured Finance," published Oct. 18, 2019).For loans with interest rates indexed to one-year LIBOR (5.4%) and one-year Treasury index (2.8%)



2668248


we applied our forward stresses, based on our criteria.

Imputed Promises Analysis

We impute the interest owed to the security holders when rating U.S. RMBS transactions wherecredit-related events can reduce interest owed to the tranches across the capital structure ratherthan an allocation of that credit-related loss to the available credit support, based on our criteria.WAC deterioration that occurs because of defaults, repurchases, or prepayments is alreadyaccounted for in our analysis or not considered credit-related; therefore, it is not considered aspart of this analysis.

This transaction provides for credit-related loan modifications and extraordinary trust expensesto reduce the net WAC, at which the transaction's bond coupons are capped. Therefore, we appliedthe approach outlined in the criteria to assess the maximum potential rating (MPR) that couldapply, based on our projected interest reduction amount (PIRA). Because this is a new issuetransaction, we did not account for any cumulative interest reduction amount.

Consistent with our criteria, we assumed that 50.0% of the loans projected to default would bemodified. We also assumed that 75.00% of the projected

modifications are interest rate modifications, with an interest rate reduction of 2.00%. Whenadded to the extraordinary trust expenses, this resulted in a maximum PIRA on the preliminaryrated certificates that is below the 4.50% threshold. We stressed extraordinary trust expenses bythe relevant extraordinary expense application factor over four years from payment periods 13 to60. Based on the results of our analysis, there was no impact on the securities' MPR.

Historically, we have observed that extraordinary trust expenses have been both minimal whenthey occur and extremely limited in pre-2009 RMBS transactions. We continue to expect theiractual occurrence in post-2009 transactions to be rare.

Operational Risk Assessment

Our criteria "Global Framework For Assessing Operational Risk In Structured FinanceTransactions," published Oct. 9, 2014, present our methodology and assumptions for assessingcertain operational risks (severity, portability, and disruption) associated with asset types and keytransaction parties (KTPs) that provide an essential service to a structured finance issuer. Asoutlined in the criteria, we cap the ratings on a transaction if we believe operational risk could leadto credit instability and affect the ratings.

As provided in the operational risk criteria, for severity risk and portability risk, there are threepossible rankings: high, moderate, or low. For disruption risk, there are four possible rankings:very high, high, moderate, or low. The rankings for each of the risks determine the MPR that can beassigned to a structured finance security for a given KTP before giving consideration to anyprovisions for a backup KTP, such as a master servicer.

According to our criteria, we rank severity and portability risk for nonprime residential mortgagecollateral as moderate and low, respectively. For GCAT 2021-NQM3, the master servicer,Nationstar Mortgage, LLC, is the KTP. We assess the disruption risk for Nationstar as low. Giventhese risk assessments, our criteria does not cap the ratings on the transaction.



2668248


Related Criteria

- Criteria | Structured Finance | General: Global Framework For Payment Structure And CashFlow Analysis Of Structured Finance Securities, Dec. 22, 2020

- Criteria | Structured Finance | General: Methodology To Derive Stressed Interest Rates InStructured Finance, Oct. 18, 2019

- Criteria | Structured Finance | Legal: U.S. Structured Finance Asset Isolation AndSpecial-Purpose Entity Criteria, May 15, 2019

- Criteria | Structured Finance | General: Counterparty Risk Framework: Methodology AndAssumptions, March 8, 2019

- Criteria | Structured Finance | General: Incorporating Sovereign Risk In Rating StructuredFinance Securities: Methodology And Assumptions, Jan. 30, 2019

- Criteria | Structured Finance | RMBS: U.S. Residential Mortgage Operational AssessmentRanking Criteria, Feb. 22, 2018

- Criteria | Structured Finance | RMBS: Methodology And Assumptions For Rating U.S. RMBSIssued 2009 And Later, Feb. 22, 2018

- Criteria | Structured Finance | RMBS: Assumptions Supplement For Methodology AndAssumptions For Rating U.S. RMBS Issued 2009 And Later, Feb. 22, 2018

- General Criteria: Methodology For Linking Long-Term And Short-Term Ratings, April 7, 2017

- Criteria | Structured Finance | General: Global Framework For Assessing Operational Risk InStructured Finance Transactions, Oct. 9, 2014

- General Criteria: Global Investment Criteria For Temporary Investments In TransactionAccounts, May 31, 2012

- General Criteria: Principles Of Credit Ratings, Feb. 16, 2011

- Criteria | Structured Finance | General: Global Methodology For Rating Interest-Only Securities,April 15, 2010

- Criteria | Structured Finance | General: Methodology For Servicer Risk Assessment, May 28,2009

Related Research

- S&P Global Ratings Publishes List Of Third-Party Due Diligence Firms Reviewed For U.S. RMBS,June 10, 2021

- Select Servicer List, June 3, 2021

- Economic Outlook U.S. Q2 2021: Let The Good Times Roll, March 24, 2021

- S&P Global Ratings Definitions, Jan. 5, 2021

- Servicer Evaluation: Nationstar Mortgage LLC, Dec. 22, 2020

- Can COVID-19 Cause A Cash Crunch For Certain U.S. RMBS?, Aug. 21, 2020

- Non-QM RMBS And COVID-19: Locking Down States' Exposure, June 1, 2020



2668248


- Guidance: Methodology And Assumptions For Rating U.S. RMBS Issued 2009 And Later, April17, 2020

- S&P Global Ratings Is Assessing The Impact Of COVID-19 On Mortgage Market Outlooks ForGlobal RMBS, April 17, 2020

- Servicer Evaluation: Shellpoint Mortgage Servicing, April 1, 2020

- U.S. Residential Mortgage Input File Format For LEVELS, March 6, 2020

- Credit Rating Model: LEVELS Model For U.S. Residential Mortgage Loans, Aug. 5, 2019

- Credit Rating Model: Intex RMBS Cash Flow Model, April 7, 2017

- Global Structured Finance Scenario and Sensitivity Analysis 2016: The Effects of The Top FiveMacroeconomic Factors, Dec. 16, 2016

- Standard & Poor's Comfortable With SFIG Draft Proposal Regarding TRID Due Diligence, April25, 2016

- Older RMBS Transactions Face Increased Tail Risk As Their Pools Shrink, Aug. 9, 2012



2668248


S&P may receive compensation for its ratings and certain credit-related analyses, normally from issuers or underwriters of securities or from obligors.S&P reserves the right to disseminate its opinions and analyses. S&P's public ratings and analyses are made available on its Web sites,www.standardandpoors.com (free of charge), and www.ratingsdirect.com and www.globalcreditportal.com (subscription), and may be distributedthrough other means, including via S&P publications and third-party redistributors. Additional information about our ratings fees is available atwww.standardandpoors.com/usratingsfees.

S&P keeps certain activities of its business units separate from each other in order to preserve the independence and objectivity of their respectiveactivities. As a result, certain business units of S&P may have information that is not available to other S&P business units. S&P has establishedpolicies and procedures to maintain the confidentiality of certain non-public information received in connection with each analytical process.

Credit-related and other analyses, including ratings, and statements in the Content are statements of opinion as of the date they are expressed andnot statements of fact. S&P's opinions, analyses and rating acknowledgment decisions (described below) are not recommendations to purchase,hold, or sell any securities or to make any investment decisions, and do not address the suitability of any security. S&P assumes no obligation toupdate the Content following publication in any form or format. The Content should not be relied on and is not a substitute for the skill, judgment andexperience of the user, its management, employees, advisors and/or clients when making investment and other business decisions. S&P does not actas a fiduciary or an investment advisor except where registered as such. While S&P has obtained information from sources it believes to be reliable,S&P does not perform an audit and undertakes no duty of due diligence or independent verification of any information it receives. Rating-relatedpublications may be published for a variety of reasons that are not necessarily dependent on action by rating committees, including, but not limitedto, the publication of a periodic update on a credit rating and related analyses.

To the extent that regulatory authorities allow a rating agency to acknowledge in one jurisdiction a rating issued in another jurisdiction for certainregulatory purposes, S&P reserves the right to assign, withdraw or suspend such acknowledgment at any time and in its sole discretion. S&P Partiesdisclaim any duty whatsoever arising out of the assignment, withdrawal or suspension of an acknowledgment as well as any liability for any damagealleged to have been suffered on account thereof.

Copyright © 2021 Standard & Poor's Financial Services LLC. All rights reserved.

No content (including ratings, credit-related analyses and data, valuations, model, software or other application or output therefrom) or any partthereof (Content) may be modified, reverse engineered, reproduced or distributed in any form by any means, or stored in a database or retrievalsystem, without the prior written permission of Standard & Poor's Financial Services LLC or its affiliates (collectively, S&P). The Content shall not beused for any unlawful or unauthorized purposes. S&P and any third-party providers, as well as their directors, officers, shareholders, employees oragents (collectively S&P Parties) do not guarantee the accuracy, completeness, timeliness or availability of the Content. S&P Parties are notresponsible for any errors or omissions (negligent or otherwise), regardless of the cause, for the results obtained from the use of the Content, or forthe security or maintenance of any data input by the user. The Content is provided on an “as is” basis. S&P PARTIES DISCLAIM ANY AND ALL EXPRESSOR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, ANY WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE ORUSE, FREEDOM FROM BUGS, SOFTWARE ERRORS OR DEFECTS, THAT THE CONTENT'S FUNCTIONING WILL BE UNINTERRUPTED OR THAT THECONTENT WILL OPERATE WITH ANY SOFTWARE OR HARDWARE CONFIGURATION. In no event shall S&P Parties be liable to any party for any direct,indirect, incidental, exemplary, compensatory, punitive, special or consequential damages, costs, expenses, legal fees, or losses (including, withoutlimitation, lost income or lost profits and opportunity costs or losses caused by negligence) in connection with any use of the Content even if advisedof the possibility of such damages.

Standard & Poor’s | Research | June 16, 2021 33

2668248


GCAT 2021-NQM3

Documents

Transcript of GCAT 2021-NQM3