This criteria article supplements our global methodology for assessing the credit quality of pools of residential loans and the payment structure and cash flow mechanics of transactions backed by these loans. It applies to residential mortgage-backed securities (RMBS) and covered bonds we rate in the Asia-Pacific (APAC) region. For details about the initial publication of this criteria article, see "Two Sector and Industry Variables Reports Archived; Contents Moved To New Asia-Pacific RMBS Criteria Supplement," published April 4, 2024.

This article contains sector and industry variables and associated details for Australia, China, Japan, Korea, New Zealand, and Singapore.

OVERVIEW

This criteria supplement is intended to be read in conjunction with S&P Global Ratings' criteria "Global Methodology And Assumptions: Assessing Pools Of Residential Loans," published Jan. 25, 2019 (hereafter referred to as the "residential mortgage criteria") and, where relevant, "Global Framework For Payment Structure And Cash Flow Analysis Of Structured Finance Securities," Dec. 22, 2020 (hereafter referred to as the "cash flow criteria"). It provides additional information and further details on the application of the criteria to selected jurisdictions in APAC. The criteria supplement is effective in Australia, New Zealand, and Singapore as of June 28, 2024.

SECTOR AND INDUSTRY VARIABLES

The sector and industry variables and associated details in this criteria supplement are expected to be periodically updated and republished as market conditions warrant. Other relevant considerations in this criteria supplement--including articulating how we may apply specific aspects of the criteria or additional information on non-fundamental factors that analysts may consider in the application of the criteria--may also be updated over time.

CRITERIA SUPPLEMENT: AUSTRALIA

Variables Used As An Input In the Credit Analysis Of Mortgage Pools

Australia mortgage market assessment (MMA)

The MMA for Australia is "low risk" ('2' on a scale of '1' to '6'). This is based on an economic risk score of '2', a mortgage industry risk score of '2', and our assessment of recourse available to lenders in the market. As per our MMA methodology framework (see Appendix I and table 5 of our residential mortgage criteria), the MMA for Australia reflects our view of:

• The risks in the Australian banking system;
• Australia's historical unemployment sensitivity to changes in economic output, combined with a relatively supportive social welfare system; and
• Borrowers' willingness to pay in low/negative equity scenarios due to full-recourse.

Australia archetypal pool

We define the archetypal pool for Australia as follows:

Table 1

Australia--archetypal pool
Characteristics by type	Archetypal features
Pool
Pool size	At least 250 loans at issuance.
Originator	No adjustment factor related to the quality of the lender's underwriting or for features not specifically covered by other adjustments.
Geographic distribution	Diversified nationally.
Borrower
Borrower type	Borrower is a private individual and not a first-time buyer.
Employment type	Pay-as-you-go (PAYG) full time and part time.
Performance status	Not delinquent.
Borrower credit history	No adverse credit history.
Affordability	Lender has assessed and fully verified the borrower's income.
Citizenship	Australian resident.
Loan
Currency	Australian dollar
Seasoning	Up to 60 months.
Loan amortization profile	Fully amortizing.
Loan product	No payment shock feature.
Loan purpose	Purchase or to refinance the balance on an existing loan (where the lender has fully re-underwritten the loan) of a residential property for owner occupation.
Loan-to-value (LTV)	75% (calculated weighting the original LTV and current indexed LTV in an 80/20 ratio).
Net surplus ratio (NSR) or Debt-to-income (DTI)	Lender has assessed the borrower's affordability commensurate with market standards.
Security	First-lien mortgage on the property, or crown leaseholds with a lease term of at least 15 years longer than the loan term.
Interest rate	Fixed-rate or floating-rate loans.
Term to maturity	>=30 and <31 years.
Property
Property type	Residential.
Occupancy status	Owner occupied.
Valuation method	Full valuation (or appraisal) of mortgaged property by registered valuers at the time of loan approval.
Valuation amount	Up to A$2 million.

Australia 'AAA' and 'B' foreclosure frequency anchors

Based on an MMA of "low risk" ('2'), we set Australia's 'AAA' foreclosure frequency anchor for the archetypal pool at 10%, and the current 'B' foreclosure frequency assumption at 1.1%.

The 1.1% 'B' foreclosure frequency assumption reflects our assessment of historical performance and our expectations for future performance over the medium term, given likely macroeconomic conditions. In our analysis, we consider default expectations across mortgage portfolios and the performance of outstanding Australian RMBS transactions that we rate, in both cases adjusted for seasoning.

Australia archetypal foreclosure frequency anchors

The following table shows the archetypal foreclosure frequency anchors.

Table 2

Archetypal foreclosure frequency anchors
Rating level*	Archetypal foreclosure frequency (%)
AAA	10.0
AA	7.5
A	5.0
BBB	3.2
BB	2.1
B	1.1
*Assumptions for intermediate rating levels are interpolated.

Variables Considered When Calculating the Weighted-Average Foreclosure Frequency (WAFF) Of A Mortgage Pool

Australia foreclosure frequency adjustment factors for variations from the archetypal pool

Table 3

Foreclosure Frequency Adjustments
Factor	Adjustment to foreclosure frequency
Loan to value (LTV)	Type 2 LTV curve (full-recourse market). The LTV is calculated weighting 80% of the original LTV (OLTV) and 20% of the current indexed LTV (CLTV).
Loan affordability	0.95x if lender assessment more conservative than industry peers, and 1.10x if lender assessment is below industry peers.
Seasoning (adjustment factors for loan seasoning)	0.75x for seasoning >5 and <=6 years; 0.70x for seasoning >6 and <=7 years; 0.65x for seasoning >7 and <=8 years; 0.60x for seasoning >8 and <=9 years; 0.55x for seasoning >9 and <=10 years; and 0.50x for seasoning >10 years. The adjustment applies only to loans that are current or in arrears for up to 30 days. The adjustment does not apply to loans with bullet, balloon, interest-only, or negative amortization features during their period of non-amortization.
Employment status	1.5x for non-pay-as-you-go (PAYG) full time & part time (including PAYG-casual, commission-based, pension, over 65, and retired); 1.5x for self-employed of <=2 years; 1.2x for self-employed of >2 years to <=5 years; and 4.0x for unemployed
Borrower occupancy status	1.1x for investment loans or second homes
Loan purpose	1.2x for refinance with debt consolidation, or refinance with cash out/equity withdrawal.
Payment shock (Loan product type)	1.2x for teaser rate, during and up to six months after end of teaser rate period; 3.0x for balloon loans (partial amortizing with a residual payment), or bullet loans (interest only until maturity), or negative amortization loans; and 1.1x-2.0x for partially amortizing loans (interest only (IO) for a period, then revert to fully amortizing) as specified in table 4.
Arrears	2.5x for loans currently 30-59 days delinquent; 5.0x for loans currently 60-89 days delinquent; and 100% foreclosure frequency for loans currently 90 days or more delinquent.
Residency status	1.5x for non-Australian residents
Credit history and arrears history for non-conforming borrowers	2.5x for borrowers with one adverse credit history event; and 3.0x for borrowers with two or more adverse credit history events.   The number of adverse credit history events is counted in reference to the number of credit events, not the number of defaults and judgements, within the past five years of A$1,500 or higher (*).   1.1x for borrowers with two events of arrears in the past 12 months; 1.2x for borrowers with three events of arrears in the past 12 months; 1.5x for borrowers with four events of arrears in the past 12 months; and 2.0x for borrowers with five or more events of arrears in the past 12 months. For borrowers with both adverse credit history and arrears history, only the credit history adjustment factor applies. For borrowers with no adverse credit history but who have arrears history, only the arrears history adjustment factor applies.
Loan term	0.7x for loan term <30 years; and 1.2x for loan term >=31 years. This adjustment factor does not apply to loans with IO periods, bullet loans, negative amortizing loans, or balloon loans.
Geographic concentration	1.2x applied to the exposure in excess of the state concentration limits; 1.5x applied to the exposure in excess of nonmetropolitan concentration limits; and 1.5x applied to the exposure in excess of postcode concentration limits. See "Geographic concentration" section below for concentration limits.
Originator adjustment	Typically 0.7x-1.3x or higher.
Income verification for self-employed and low documentation loans	1.5x for no credible source; 1.4x for one credible source; 1.35x for two credible sources; 1.3x for three credible sources; 1.25x for four credible sources; and 1.0x if tax returns are included. Adjustment factor reduces over a six-year period as specified in table 8.
First-time buyer	1.1x to first-time borrower with no credit history, until a payment history of at least 18 months has been established, upon which the adjustment factor is no longer applied if the loan is less than 30 days in arrears.
Self-managed superannuation fund (SMSF)	1.25x
Redraws and further advances	Typically 1.05x for redraw only or for further advance only, 1.1x for both redraw and further advance. The size of these adjustment factors is based on the capacity for redraw and further advances in a pool or transaction based on loan characteristics and documented restrictions on allowable further advances. Higher adjustment may apply the if pool or transaction structure has greater capacity for redraws or further advances.
*We consider a credit event to be one event in a borrower's life that may drive a number of events (such as one loss of employment or ill health) that leads to a borrower not being able to meet payment obligations.

Loan to value (LTV)

LTV is calculated in a three-stage process.

• Stage 1: The original LTV (OLTV) is calculated using the original loan balance and the property valuation at the time of loan origination.
• Stage 2: The current LTV (CLTV) is calculated using the loan balance as of the portfolio cut-off date and the current indexed property value. The CLTV calculation incorporates any applicable valuation haircuts.
• Stage 3: The LTV is calculated by taking 80% of the OLTV and 20% of the CLTV. We may also consider the maximum drawable balance or further advance if applicable in our analysis.

The OLTV and CLTV are based on the full loan balances secured on the property, including loan parts outside the asset pool and prior- and second-ranking balances, if any. The use of an 80/20 combination of OLTV and CLTV maintains comparability to other similar jurisdictions under our global framework.

A loan's CLTV is the current balance of a loan, reflecting the actual principal pay down, divided by the indexed initial value.

The continuous function in the equation below defines the adjustment factors in chart 1.

These figures are rounded to four decimal places for ease of readability.

Loan affordability

To assess serviceability, most Australian lenders adopt a net-surplus ratio (NSR) or an equivalent derivation measure (e.g., net disposable income or dollar net surplus measure), with some supplementing NSR with debt-to-income (DTI) and or loan-to-income (LTI) ratios. Although the various approaches are relatively uniform, the composition and derivation of income and expenses, interest rate buffer and/or floor, and the maximum or minimum limits (whichever is applicable) could vary across originators. Hence any meaningful deviation from market standards by originators in their assessment of affordability could attract a positive or negative foreclosure frequency adjustment.

Payment shock (loan product type)

Although interest-only loans that revert to fully amortizing loans after the interest-only period (IO-term) expires somewhat mitigate refinancing risk, the interest-only feature can create a payment shock when the payments revert to fully amortizing over the remaining term of the loan (PI-term). The amount of the adjustment depends on the degree of payment shock in each case and is a product of the IO-term-related adjustment factor and PI-term-related adjustment factor in table 4. This reflects our view that the longer the IO-term is relative to the PI-term, the higher the risk of payment shock-related default.

Table 4

Adjustments For Interest-Only (IO), Then Reverting To Fully Amortizing (PI) Loans
IO term (years)	<=5	>5 to <=10	>10 to <=15	>15 to <=20	>20 to <=25
Adjustment factor	1.1x	1.25x	1.5x	1.75x	2.0x
PI term (years)	<3	=3 to <5	=5 to <10	=10 to <15	=15 to 30
Adjustment factor	1.75x	1.5x	1.25x	1.1x	1.0x

Originator adjustment

The calculation of foreclosure frequency includes an originator adjustment. Specific examples of the factors considered in determining the originator adjustment for a pool of Australian residential loans typically include, but are not limited to:

• The quality of origination and underwriting. Our review assesses a range of issues including the business and lending strategy; the governance and management structure; risk-management, compliance, quality assurance (including data quality and internal control framework); policies, procedures and training programs; the role of mortgage originators and brokers in the origination and underwriting process; the underwriting standards and quality of credit review (including whether credit checks have been obtained, verification of savings history, quality of valuation policy); the frequency of exceptions to established underwriting guidelines; general insurance coverage over security properties; and the role of any mortgage insurer in the underwriting process (if applicable).
• The servicing quality. This includes a review of the loan servicing philosophy and experience, the experience in servicing assets of the type to be securitized, the quality of the loan-servicing platform, the financial strength of the servicer, the experience of management and staff, the role of mortgage originators and brokers, the risk management, quality assurance, compliance and internal control framework, arrears management and recoveries, the historical delinquency performance, the transaction processing and cash management, and investor reporting.
• The historical performance of an originator's loans relative to other originators.
• Where insufficient data are available on borrowers' employment status, S&P Global Ratings typically assumes a percentage of the pool is to self-employed borrowers, generally 25% of all full-documentation loans, and all low- and no-documentation loans. We may assume a different level to this for full-documentation loans where line-by-line data is not available, if we are of the opinion that the characteristics of the originator or the borrowers would suggest a different percentage is more appropriate. We apply a stress of 100% for low- and no-documentation loans because such loans are almost always provided to self-employed borrowers.
• Where the borrower's loan purpose is for refinancing and there is insufficient data available on whether the refinancing is with or without debt consolidation or cash out/equity withdrawal, we may assume a portion of the pool is used for refinancing with debt consolidation or cash out/equity withdrawal. Such an assessment considers the lender's underwriting policies and practices, the lender's portfolio parameters, and industry trends and averages.
• Where insufficient data is available on the borrower's first-time home buyer status, we typically assume a percentage of the pool is made to first-time home buyers, generally within a range of 10%-30%. We may assume a higher level than this where line-by-line data is not available if we are of the opinion that the characteristics of the originator or the borrowers would suggest a higher percentage is more appropriate.

Geographic concentration

We apply adjustment factors to the exposure in excess of the specific thresholds for each state, nonmetropolitan, and postcode shown in the table below.

Table 5

Geographic Concentration Limits By State
State	Limit (%)
New South Wales	<=60
Victoria	<=50
Queensland	<=40
Western Australia	<=25
South Australia	<=25
Australian Capital Territory	<=5
Tasmania	<=5
Northern Territory	<=5

Table 6

Geographic Concentration Limits By Nonmetropolitan And Postcode
State	Limit (%)
Nonmetropolitan	<=10
Postcode	<=2

Classifications of postcodes in Australia are shown in the table below.

Table 7

Australian postcode classifications
Postcode	State	Location
0200-0799	Australian Capital Territory	Nonmetro
0800-0820	Northern Territory	Metro
0821-0827	Northern Territory	Nonmetro
0828-0832	Northern Territory	Metro
0833-0999	Northern Territory	Nonmetro
1000-1920	New South Wales	Metro
1921-1999	New South Wales	Nonmetro
2000-2005	New South Wales	Metro
2006-2308	New South Wales	Metro
2309-2499	New South Wales	Nonmetro
2500-2534	New South Wales	Metro
2535-2554	New South Wales	Nonmetro
2555-2574	New South Wales	Metro
2575-2599	New South Wales	Nonmetro
2600-2617	Australian Capital Territory	Metro
2618-2744	New South Wales	Nonmetro
2745-2786	New South Wales	Metro
2787-2899	New South Wales	Nonmetro
2900-2920	Australian Capital Territory	Metro
2921-2999	New South Wales	Nonmetro
3000-3010	Victoria	Metro
3011-3232	Victoria	Metro
3233-3234	Victoria	Nonmetro
3235	Victoria	Metro
3236-3239	Victoria	Nonmetro
3240-3241	Victoria	Metro
3242-3320	Victoria	Nonmetro
3321	Victoria	Metro
3322-3327	Victoria	Nonmetro
3328-3340	Victoria	Metro
3341-3426	Victoria	Nonmetro
3427-3441	Victoria	Metro
3442-3749	Victoria	Nonmetro
3750-3815	Victoria	Metro
3816-3909	Victoria	Nonmetro
3910-3920	Victoria	Metro
3921-3925	Victoria	Nonmetro
3926-3944	Victoria	Metro
3945-3971	Victoria	Nonmetro
3972-3978	Victoria	Metro
3979	Victoria	Nonmetro
3980-3983	Victoria	Metro
3984-3999	Victoria	Nonmetro
4000-4004	Queensland	Metro
4005-4228	Queensland	Metro
4229-4269	Queensland	Nonmetro
4270-4313	Queensland	Metro
4314-4339	Queensland	Nonmetro
4340-4342	Queensland	Metro
4343-4345	Queensland	Nonmetro
4346	Queensland	Metro
4347-4499	Queensland	Nonmetro
4500-4575	Queensland	Metro
4576-4999	Queensland	Nonmetro
5000-5005	South Australia	Metro
5006-5199	South Australia	Metro
5200-5799	South Australia	Nonmetro
5800-5999	South Australia	Metro
6000-6004	Western Australia	Metro
6005-6214	Western Australia	Metro
6215-6799	Western Australia	Nonmetro
6800-6999	Western Australia	Metro
7000-7003	Tasmania	Metro
7004-7199	Tasmania	Metro
7200-7799	Tasmania	Nonmetro
7800-7899	Tasmania	Metro
7900-7999	Tasmania	Nonmetro
8000-8399	Victoria	Metro
8400-8899	Victoria	Metro
8900-8999	Victoria	Nonmetro
9000-9299	Queensland	Metro
9300-9399	Queensland	Nonmetro
9400-9596	Queensland	Metro
9597	Queensland	Nonmetro

Income verification

A borrower's demonstrated loan repayment outweighs the level of income and asset verification at the time of loan origination, and the initial level of income and asset verification becomes less indicative of likelihood of default over time. To account for this, these criteria reduce the adjustment factors over a six-year period using the equation below, which incorporates a seasoning adjustment from 100% to 0% (see table 8) over the same period. Effectively, the documentation-related adjustment factor becomes neutral after six years.

1 + (initial documentation type adjustment – 1) x seasoning adjustment

Table 8

Percentage Of Documentation Adjustment Factors Applied By Loan Seasoning
Loan Seasoning	<=12	(12-24]	(24-36]	(36-48]	(48-60]	(60-72]	>72
Seasoning adjustment	100%	85%	80%	55%	35%	15%	0%

Variables Considered When Calculating The Weighted-Average Loss Severity (WALS) Of A Mortgage Pool

Table 9

Loss Severity (WALS)
Factor	Adjustment to loss severity calculation
Valuation haircut	5% if not a full valuation.   If pool is concentrated to non-Australian residents, 20% if based on full valuation, 25% if based on all other types of valuation.
Property indexation	Based on residential property price data published by organizations or institutions we deem reputable.   We reflect 50% of index appreciation and 100% of index depreciation.
Over/undervaluation	See the "Assessment of property of over-/under-valuation" section of our residential mortgage criteria.
Property value	1.2x for property values of more than A$2 million and less than or equal to A$3 million; 1.225x for property values of more than A$3 million and less than or equal to A$4 million; 1.25x for property values of more than A$4 million and less than or equal to A$5 million; 1.275x for property values of more than A$5 million and less than or equal to A$6 million; and 1.3x for property values of more than A$6 million.
Foreclosure costs	A$10,000 fixed cost. 5% variable as a percentage of index-adjusted, post-repo, post valuation haircut if any, and post jumbo-adjusted valuation.
Foreclosure timeline/period	For property values of less than or equal to A$2 million, 12 months if located in metropolitan area, and 18 months if located in non-metropolitan area. For property values of greater than A$2 million, 18 months if located in metropolitan area, and 24 months if located in non-metropolitan area.
Accrued and unpaid interest	No adjustment where cash flow analysis is performed; otherwise included in the loss severity calculation based on current interest rate through the foreclosure period.
The MVD of a repossessed property (Repo MVD) is capped at 75%.

Valuation haircut

We apply a 5% haircut to valuation if any valuation other than full valuation is relied upon at point of loan underwriting.

If a pool is concentrated to non-Australian residents, we apply a valuation haircut of 20% if the valuation is based on full valuation, and 25% haircut if it is based on all other types of valuation. The higher haircut for pools concentrated to non-Australian residents reflects our view that because of the restrictions on nonresidents purchasing established Australian dwellings, in a foreclosure scenario, the properties purchased by nonresidents could only be sold to Australian residents. In such a situation, there can be difference in price that one cohort (nonresidents) are willing to pay versus the market value when such properties are sold to another cohort (Australian residents).

Accrued and unpaid interest

Where the rating analysis includes cash flow modelling, the impact on a transaction's credit enhancement from accrued and unpaid interest on defaulted loans during the foreclosure period is incorporated into our cash flow analysis and is therefore not included in the loss severity calculation. Where the rating analysis does not include cash flow modelling, the accrued and unpaid interest is included in the loss severity calculation. In both situations, stressed interest rate assumptions is used.

Further Considerations

Lenders mortgage insurance

Some residential loans in Australia are insured for loss by lenders mortgage insurance. In such cases, we apply the methodology outlined in "Methodology For Assessing Mortgage Insurance And Similar Guarantees And Supports In Structured And Public Sector Finance And Covered Bonds," Dec. 7, 2014.

Borrower concentration analysis

In pools where concentrations to borrowers are evident, we may apply an additional minimum loss projection when determining the expected loss for a pool at the relevant rating level.

Pools may be concentrated to borrowers with large loan sizes relative to total pool size at close, or become concentrated as pools amortize. A loss on one or a small number of loans in a concentrated pool may have a disproportionate impact on credit enhancement, because the loss on specific large loans might be higher than the absolute level of credit enhancement.

To address the potential that greater losses could result if the loans with higher balances were to default, the criteria use the expected loss on the largest loan balances.

We typically apply the additional minimum loss projection when the top-10 borrower exposures exceed 10% of the total current pool balance.

The additional minimum loss projection estimations at each rating category are based on a certain number of loans defaulting and liquidating as outlined in table 10. To address the potential that greater losses could ensue if the loans with higher balances were to default, the criteria use the largest liquidation amounts for each rating category.

The minimum loss projection estimations at each rating level are based on a certain number of loans defaulting and liquidating as outlined in table 10, and are the higher of that derived using the RMBS credit criteria outlined above and that calculated as an additional minimum loss projection in accordance with this subsection.

Table 10

Rating Category Minimum Loss Projections For Concentrated Pools
Maximum potential rating	Minimum loss parameters
AAA	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'AAA' loss severity and the weighted-average loss severity at 'AAA' for the pool, plus the next nine largest loss exposures at the 'AAA' loss severity
AA category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'AA' loss severity and the weighted-average loss severity at 'AA' for the pool, plus the next seven largest loss exposures at the 'AA' loss severity
A category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'A' loss severity and the weighted-average loss severity at 'A' for the pool, plus the next five largest loss exposures at the 'A' loss severity
BBB category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'BBB' loss severity and the weighted-average loss severity at 'BBB' for the pool, plus the next three largest loss exposures at the 'BBB' loss severity
BB category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'BB' loss severity and the weighted-average loss severity at 'BB' for the pool, plus the next largest loss exposure at the 'BB' loss severity
B category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'B' loss severity, and the weighted-average loss severity at 'B' for the pool

We may also adapt the above approach to assess the features (such as arrears status) of particular loans that account for the top X loans (X being the number of loans as determined in accordance with table 10), to capture any potential for loss on those loans to be greater than that determined under the calculation outlined in table 10.

CRITERIA SUPPLEMENT: CHINA

Variables Used As An Input In Our Calculation Of The Foreclosure Frequency

China Mortgage Market Assessment (MMA)

Our MMA in a given jurisdiction consists of an assessment on a six-point scale ranging from "very low risk" to "extremely high risk".

Table 11

Mortgage market assessment
China MMA	Intermediate risk
Sub factors
Economic risk	High risk
Industry risk	Intermediate risk
Lender - recourse assessment	Full recourse

As per our MMA methodology Framework (see Appendix I and table 5 of our residential mortgage criteria), the MMA for China reflects our view of:

• The risks in the Chinese banking system;
• Relative to the overall Chinese banking system, stronger household financials and a moderate risk appetite for mortgage loans lending, compared with corporate loans and other household loans;
• On the one hand, low unemployment sensitivity to changes in economic output in China and high household savings rate reduce credit risk in the mortgage market. On the other hand, a relatively less-supportive social welfare system, compared with some of the countries with well-established systems, increases the risk; and
• Chinese mortgage loans are full recourse, meaning that borrowers remain incentivized to pay in low/negative equity scenarios.

China archetypal foreclosure frequency anchors

When applying paragraphs 15, and 26, and table 1 of our residential mortgage criteria, based on an MMA of "intermediate risk" ('3'), we set China's foreclosure frequency assumption as indicated in table 12 below. As per paragraphs 23-25 of our residential mortgage criteria, the 'B' foreclosure frequency assumption reflects our assessment of historical performance and our expectations for future performance over the medium term, given our outlook on macroeconomic conditions. In our analysis, we considered default expectations across mortgage portfolios and the performance of outstanding Chinese RMBS transactions that we rate, in both cases adjusted for seasoning to reflect expected lifetime defaults.

Table 12

Foreclosure Frequency Anchors
Rating level*	Archetypal foreclosure frequency (%)
AAA	15.0
AA	11.5
A	8.0
BBB	5.5
BB	4.0
B	2.5
*Assumptions for intermediate rating levels are interpolated.

China archetypal pool

We define the archetypal pool for China as follows:

Table 13

China--Archetypal Pool
Characteristics by type	Archetypal features
Pool
Pool size	At least 250 loans at issuance.
Originator	No adjustment factor related to the quality of the lender's underwriting or for features not specifically covered by other adjustments.
Geographic distribution	Diversified nationally.
Borrower
Borrower type	Borrower is a private individual.
Citizenship	Chinese citizen.
Employment type	Not self-employed or unemployed.
Performance status	Not delinquent.
Borrower credit history	No adverse credit history.
Affordability	Lender has assessed the borrower's income.
Loan
Currency/denomination	RMB
Seasoning	Up to 60 months.
Loan amortization profile	Fully amortizing.
Loan product	Full-recourse loans with no payment shock or interest-only to maturity feature.
Interest rate	Floating-rate loans.
Security	First-lien mortgage on the property.
Loan purpose	Purchase or to refinance the balance on an existing loan (where the lender has fully re-underwritten the loan) of a residential property for owner occupation.
Loan-to-value (LTV)	70% (calculated weighting the original LTV and current indexed LTV in an 80:20 ratio).
Debt-to-income (DTI)	35% mortgage loan only DTI.
Property
Property type	Residential.
Occupancy status	Owner occupied and primary residence.
Valuation method	Full valuations on the mortgaged property from a real estate appraiser.
Valuation amount	Up to the applicable jumbo valuation threshold specified in table 16.
Mortgage registration	Completed.

Variables Considered When Calculating The Foreclosure Frequency Of A Mortgage Pool

For each characteristic and attribute that differs from the archetype at a pool, loan, or borrower level, the criteria apply a corresponding adjustment to the foreclosure frequency. The following table shows the applicable adjustments.

Table 14

Adjustment Factors Applied When Calculating The Foreclosure Frequency Of A Mortgage Pool
Originator adjustment	Typically 0.7x-1.3x or higher applied at pool level. For country-specific originator adjustment factors, see "Originator Adjustment" paragraphs below. In particular, we may apply an originator adjustment to capture missing loan-level data for self-employed borrowers or loans that are backed by second homes, or investment properties where the loan is originated against the borrower's income. In sizing this adjustment, for pools that we consider to be representative of the overall mortgage market, we typically assume 45% of loans are self-employed borrowers, and 30% of loans are backed by second homes.
LTV	Type 2 LTV curve (see chart 2 below).
Combined LTV definition	The LTV is calculated weighting: 80% of the OLTV and 20% of the CLTV.
Loan affordability: Debt-to-income (DTI)	0.8x if DTI <= 27%; 1.0x if DTI = 35%; and 1.5x if DTI >= 55%. This adjustment is a continuous function (see chart 3 below). For borrowers without proven-income, higher of DTI and income documentation adjustment factor is applied. Only non-investment loans will be subject to the DTI adjustment.
Income documentation	Typically 1.5x adjustment for self-certification and missing data; removed gradually when the loan is performing or in arrears for less than 30 days, and seasoning exceeds 12 months as follows: <=12: 100%; 12-24: 85%; 24-36: 80%; 36-48: 55%; 48-60: 35%; 60-72: 15%; and >72: 0. For borrowers without proven income, higher of DTI and income documentation adjustment factor is applied. Only non-investment loans will be subject to the income documentation adjustment.
Seasoning (adjustment factors for loan seasoning)	1.0x for seasoning <=5 years; 0.75x for seasoning >5 and <=6 years; 0.70x for seasoning >6 and <=7 years; 0.65x for seasoning >7 and <=8 years; 0.60x for seasoning >8 and <=9 years; 0.55x for seasoning >9 and <=10 years; and 0.50x for seasoning >10 years.   Factor applies only to loans that are current or in arrears less than 30 days.
Employment status	1.0x for regular salaried employee; 1.25x for self-employed person and irregular salaried employee or contractor; 1.5x for pensioner and retired borrowers; and 1.75x for unemployed person or other employment status.   Only non-investment loans will be subject to the employment status adjustment.
Second and subsequent lien loans	1.3x: Where we consider that the loan does not have significant risk layering, or where the second and subsequent lien was not taken out to consolidate debt (is akin to a further advance), and where there is data relating to the senior lien holder; 1.5x: Where the borrower is using the second and subsequent lien for consumption or consolidation of debt and where we consider that there is risk layering; and 1.7x: Where there is insufficient data to back up other second and subsequent lien adjustments.
Borrower occupancy status: investment property (buy-to-let)	1.7x for loans underwritten to non-borrower income, such as rental income from the investment properties.
Borrower occupancy status: owner-occupied (second home)	1.3x for second home and loans identified with investment purpose underwritten to borrower income if any.
Loan purpose	1.1x for refinancing unless the lender has undertaken a full re-underwriting procedure; and 1.2x for debt consolidation, equity release/cash out loans and other loan purpose.   The adjustment does not apply if the loan is also second or subsequent lien.
Loan product type: payment shock and interest only	1.0x for fixed rate, fixed reset, and floating rate; 1.2x for fixed to floating rate, removed six months after the end of the fixed-rate period; 1.2x for teaser rate, removed six months after the end of teaser-rate period; 1.2x for IO loans, removed six months after the end of interest-only period; 1.2x for incremental repayment; 1.3x for balloon loans; 1.5x for bullet loans (interest-only to maturity); and 2.0x for negative amortization.
Arrears	2.5x for loans currently 30-59 days delinquent; 5.0x for loans currently 60-89 days delinquent; and 100% foreclosure frequency for 90 days or more delinquent.
Reperforming loans	We typically apply adjustments for reperforming loans when a portfolio contains a material portion of such loans as follows: Months since last 90+ days in arrears or restructure dates <=12: 2.0x; >12<=24: 1.5x; >24<=36: 1.2x; and >36: No adjustment.
Nonresidential use loans / commercial borrowers	1.5x-2.0x multiple for commercial/ mixed-use property; and 2.0x multiple for commercial borrowers. Limit of 40% of the pool at issuance.
Geographic concentration	1.2x adjustment applied to the excess above the city concentration thresholds (% of pool balance in table 15 below).
Residency status	Up to 2.5x adjustment factor for non-Chinese citizens.
Construction Home (Forward-delivery houses without mortgage registration)	2x for forward-delivery houses without mortgage registration as of pool cut-off date. We may adjust this multiple based on originator-specific performance data and underwriting criteria. Adjustment factor is removed when the loan is seasoned for three years or above since origination, or when we are informed the mortgage registration has been completed. This approach applies only for pool concentrations up to 10%. If the exposure exceeds 10%, we would give no credit to the excess.

Table 15

Concentration Limits For Geographic Concentration Adjustment Factor
City	Concentration limit (%)
Beijing	4.0
Shanghai	4.0
Guangdong Guangzhou	4.0
Guangdong Shenzhen	4.0
Chongqing	4.0
Sichuan Chengdu	2.0
Tianjin	2.0
Hebei Baoding	2.0
Hubei Wuhan	2.0
Hebei Shijiazhuang	2.0
Heilongjiang Harbin	2.0
Jiangsu Suzhou	2.0
Shandong Linyi	2.0
Henan Zhengzhou	2.0
Zhejiang Hangzhou	2.0
Shandong Qingdao	2.0
Fujian Quanzhou	2.0
Jiangsu Nanjing	2.0
Zhejiang Ningbo	2.0
Hunan Changsha	2.0
Others	0.5

Details Of Further Considerations When Calculating The Foreclosure Frequency Of A Chinese Mortgage Pool

Originator adjustment

Specific examples of the factors considered in determining the originator adjustment for a pool of Chinese residential loans typically include, but are not limited to:

• Origination and underwriting quality;
• Servicing quality;
• Originator-specific performance, to address idiosyncratic risks that are not captured in other adjustments;
• If we believe the originator's LTV calculation is materially different from market standards or we believe inclusion of second and subsequent-ranking balances will significantly increase LTV levels;
• If we believe the originator's practices on DTI calculation or requirements are significantly different from market standards, or the performance data suggests the need for additional negative or positive adjustment;
• If we believe additional risk is not sufficiently captured in loan product adjustments;
• Unavailable data on loan and borrower characteristics (for example, LTV, DTI, income verification, restructurings, defaults, prior mortgage arrears, and property valuation method);
• If originators can't provide data at the loan level for employment status, second homes or investment properties underwritten to borrower income, we apply an originator adjustment (see table 14 above). We typically size this originator adjustment by multiplying the applicable adjustment with our assumption of the proportion of loans with the relevant characteristic in the pool. We determine this assumption based on pool-level data, if available, the originator's underwriting policies, or market-level data, if representative;
• Loans to borrowers with negative credit histories;
• Recent changes in product offering, credit score process or underwriting guidelines, where the impact of which is not yet visible in performance metrics;
• Positive or negative selection not captured in other adjustments;
• Scenarios where the performance of a pool that has been sold deviates from our expectations for pools from that originator and for which a neutral originator adjustment is assumed;
• Dynamic or revolving asset pools;
• Any transaction's specific geographical concentration not captured by city concentrations;
• We may analyze data from the issuer/servicer for reperforming loans on re-default rates stratified by forbearance type to calibrate the originator adjustment for such transactions; and
• If forward-delivery houses without mortgage registration are concentrated on certain developers as of the pool cut-off date.
• Any other observed risk characteristics that are not adjusted for at a loan-by-loan level.

Loan to value (LTV)

LTV is calculated in a three-stage process. Stage 1: The OLTV is calculated using the original loan balance at the time of the latest advance, typically at loan origination and the property valuation at the time of that advance. Stage 2: The CLTV is calculated using the loan balance as of the portfolio cut-off date and the current indexed property value. The CLTV calculation incorporates any applicable valuation haircuts. Stage 3: The LTV is calculated weighting 80% of the OLTV and 20% of the CLTV. We may also consider the maximum drawable balance, further advance, and purchase price, if applicable, in our analysis.

The OLTV and CLTV are based on the full loan balances secured on the property, including loan parts outside the asset pool and prior-ranking balances. CLTV could reflect updated prior-ranking balances if data is available. Second and subsequent-ranking balances will be also included in both OLTV and CLTV if data is available.

An originator adjustment may apply if the originator's LTV calculation in their underwriting criteria is materially different from market standards or we believe the inclusion of second and subsequent-ranking balances will significantly increase the portfolio's LTV levels.

Loan affordability: Debt-to-income (DTI)

DTI is defined as principal and interest payment for mortgage loans only over income (following the definition by local regulators). Secondary income and spousal income might be aggregated. If spousal income is considered, spousal debt will also be considered.

If loan-level data is not available for measuring affordability, we make adjustments based on our assessment of the originator's practices relative to market standards via the originator adjustment at the pool level.

We may apply originator adjustments in addition to DTI adjustments on a case-by-case basis, depending on the performance data provided, or if we believe the originator's practices vary significantly from market standards.

Seasoning (adjustment factors for loan seasoning)

If corresponding data are available, the criteria may consider the seasoning of the performing relationship between the borrower and the mortgage lender of an existing mortgage loan being refinanced, instead of seasoning of the new loan that results from such refinancing. For instance, a loan may have been refinanced with a new loan that has the same term or a shorter term, and the loan balance and interest rate are the same or lower. In that case, seasoning could be based on the origination date of the original loan rather than the date of the refinancing.

Second and subsequent lien loans

We apply an adjustment factor for second and subsequent-lien mortgage loans (see table 14 above). We differentiate a variable by the data available on the other lien loan.

We may consider different adjustments on a case-by-case basis, depending on the servicers' underwriting policies, track record, and performance data provided.

To the extent a pool exhibits a concentration of third and subsequent lien loans, then we may apply a higher originator adjustment.

In all cases above, the loan purpose adjustment does not apply.

Borrower occupancy status

We apply adjustment factors for loans used to finance investment properties. The investment property classification includes second homes, investment properties with loans underwritten to borrower income, and investment properties with loans underwritten to non-borrower income, such as a property's rental income or debt-service coverage ratio (DSCR).

Typically, for investment properties with loans underwritten to non-borrower income, borrower-related adjustments such as DTI, self-certification, and employment status do not apply. We may consider different adjustments on a case-by-case basis, depending on the performance data provided.

Loan product type (payment shock and interest-only)

Where a borrower may face payment shock from two separate sources, for example, when assessing a borrower with a fixed-to-float and incremental repayment loan, we do not apply the adjustment twice, but apply the highest adjustment once.

We might apply additional adjustments based on product type through originator adjustments if we believe the risks are not sufficiently captured in abovementioned adjustments.

Re-performing loans and restructured/modified loans

We may consider different adjustments on a case-by-case basis, depending on the servicers' restructure policies, track record, and performance data provided.

We typically define a reperforming loan as a loan that has been 90 or more days past due, or restructured in the three years leading up to the analysis date, and is current as of that date.

When a reperforming arrangement is made, generally a full reassessment of the borrower's affordability capacity is also carried out. We consider this akin to a re-underwriting of the loan. Accordingly, for pools classified as reperforming, we calculated potential future seasoning credit based on the date a loan was last 90 or more days in arrears.

In addition to the original loan and borrower information provided, we may also consider updated data sourced through the restructuring process in our analysis of reperforming loans, where available, on a case-by-case basis.

In addition, as part of the analytical process, we analyze data from the issuer/servicer on re-default rates stratified by forbearance type. This analysis is used to calibrate the originator adjustment for such transactions.

Geographic concentration

If a pool has significant geographic concentration risk that we believe is not sufficiently captured, we may account for it using the originator adjustment.

Residency status

We apply a lower adjustment if the originator or servicer provides evidence that the origination process is robust and the performance of loans to non-Chinese citizens is comparable to those given to Chinese citizens.

Construction Home (Forward-delivery houses without mortgage registration as of the pool cut-off date)

The definition of forward-delivery houses is that when a building is topped out (i.e. its roof is sealed), a bank can apply for an advance registration notice of mortgage and disburse a mortgage loan based upon such notice in compliance with local regulator's policy. The time taken to complete construction for delivery typically ranges from a few months to two to three years. When the houses are completed and associated certificates become available, the bank has the priority right to register a typical mortgage on the house. In our view, mortgage loans extended against forward-delivery houses are exposed to additional construction risk and hence require additional adjustment.

We may consider different adjustments on a case-by-case basis, depending on performance data provided or the originator's underwriting criteria.

We may include further changes through an originator adjustment if forward-delivery houses without mortgage registration are concentrated in certain developers as of the pool cut-off date.

Dynamic asset pools

We consider the potential changes of credit risk over time. This applies to RMBS transactions backed by a pool whose assets change (e.g., by virtue of loan substitutions, product switches, or of a similar nature, or revolving). In determining their weighted-average foreclosure frequency (WAFF) and the weighted-average loss severity (WALS), we consider the potential increase of credit risk over time as a result of changes in pool composition. We assess possible deterioration in pool composition based on the transaction's documented asset-eligibility criteria, the history of the originator and, in particular, any observed changes in origination, underwriting, and related performance.

Variables Considered When Calculating The Loss Severity Of A Mortgage Pool

The table below shows the variables we consider when calculating the loss severity of a mortgage pool.

Table 16

Variables considered when calculating the loss severity of a Chinese mortgage pool
Property indexation	Based on nationwide housing price index (HPI) published by the Organisation for Economic Co-operation and Development (OECD) for China. We reflect 50% of the HPI appreciation (the positive changes in the index) and 100% of the HPI depreciation (the negative changes in the index).
Over/undervaluation	We apply a stress multiple of 1.2x to the Repo MVD.
Jumbo valuation threshold*	150 square meters.
Valuation haircuts	5% if the valuation is based on sales price or other valuation methods besides full appraisal, or if loan-level data is not available. We may apply a higher haircut (generally up to 15% of the valuation) for construction home(forward-delivery houses without mortgage registration as of the pool cut-off date), depending on the uncertainty of property valuations to be realized upon completion. Adjustment factor will be removed when the loan is seasoned for three years or longer since origination (the stress period could vary depending on information available to us or our observation), or when we are informed the mortgage registration has been completed.
Foreclosure costs§	RMB2,000 fixed cost and 12% variable cost. In addition to the abovementioned costs, 13% in variable costs applies to loans that can be identified as being backed by only-home properties. If only-home data can't be provided at the loan level, a further 9% in variable costs will be applied to whole pool instead. All variable costs are modeled as a percentage of the post-repo, jumbo-adjustment, commercial/mixed-use properties adjustment, construction home adjustment valuation. Only-home cost can be lowered if more supporting data can be provided and analytical finding warrants it.
Foreclosure timeline/period	24 months for tier 1 cities, namely Beijing, Shanghai, Guangzhou, and Shenzhen; and 48 months for other cities. For forward-delivery houses without mortgage registration as of the pool cut-off date, plus an additional three years (the stress period could vary depending on information available to us or our observation) to the aforementioned foreclosure timeline, reduced by seasoning until the loan is seasoned for three years or longer since origination, or when we are informed the mortgage registration has been completed.
Commercial/mixed-use properties	1.15x adjustment to the MVD.
Construction Home (Forward-delivery houses without mortgage registration as of the pool cut-off date)	1.2x multiple to the Repo MVD if originators or servicers have a prudent process to manage and monitor the conversion from advance registration notice of mortgage to mortgage registration, and without concentration on certain developers as of pool cut-off date; otherwise, 100% loss severity.   Adjustment factor will be removed when the loan is seasoned for three years or longer since origination (the stress period could vary depending on information available to us or our observation), or when we are informed the mortgage registration has been completed. We might apply additional adjustments or abovementioned adjustments could vary depending on data available to us.
*We increase MVD assumptions for jumbo valuations. We apply an adjustment of 20% on the excess above the jumbo threshold. For example, for a property with a size of 170 square meters, we apply an adjustment on the difference between 170 square meters and 150 square meters (jumbo valuation threshold). Jumbo valuation adjustment, the product of 20% x 20 square meters as a percentage of property size of 170 square meters in this example, is then deducted as a proxy of market value decline from the indexed valuation after applying the post-repossession market value decline. §The post-repo, jumbo-adjustment, commercial/mixed-use properties, construction home adjustment valuation is the property value obtained after the application of the repossession market value decline at the relevant rating as well as the jumbo valuation adjustment, commercial/mixed-use properties adjustment and construction home adjustment to the indexed valuation. **MVD--Market-value decline. The MVD of a repossessed property (Repo MVD) post all adjustments is capped at 75%.

Details Of Further Considerations When Calculating The Loss Severity Of A Chinese Mortgage Pool

Jumbo valuation threshold

Jumbo valuation threshold partially reflects the view of local regulators on non-ordinary housing, measured by property size, where unfavorable regulatory requirements might be imposed. This could lead to illiquidity and a higher market-value decline when foreclosing such large properties. Other countries typically use property value as a measure to determine jumbo valuation.

Forced-sale discount (FSD)

On a case-by-case basis, we may increase or decrease the forced-sales discount where there is sufficient information to support it being either higher or lower than envisaged using the standard calculation.

Accrued and unpaid interest

Prior-ranking balances will not accrue interest and will not result in unpaid interest to be incorporated into a loan's loss severity if a cash flow analysis is not performed.

Details Of Further Considerations When Calculating Minimum Credit Enhancement Levels In The Analysis Of A Chinese RMBS Transaction

Table 17

Minimum Credit Enhancement
Rating level	AAA	AA	A	BBB	BB	B
CE Floor*	4.00	2.50	1.50	1.00	0.50	0.35
Hard CE Floor	2.50	2.00	1.50	N/A	N/A	N/Aa
*CE Floor is applied to final CE results in cash flow analysis, instead of credit losses from the credit model. CE--Credit enhancement. N.A--Not applicable.

Variables considered in the cash flow model runs

Table 18

Variables Considered In The Cash Flow Model Runs For A Chinese Mortgage Pool
Default timing scenarios	Default timing curves: Typically, "front-loaded" and "back-loaded".   The peak level of stress in each scenario is referred to as a recession for RMBS. It corresponds to months 1-60 in both scenarios. See table 19 below.
Interest rate risk	Standard run scenarios: Typically, "Up", "Up-down", "Down" and "Down-up" (see "Methodology To Derive Stressed Interest Rates In Structured Finance," published Oct. 18, 2019). The curves vary by stress scenario. Our stressed interest rate scenarios for China are published in "Sector And Industry Variables: Methodology To Derive Stressed Interest Rates In Structured Finance," published on June 30, 2023.
Prepayment scenarios	Typically, "Low" and "High". See table 20 below.
Senior fees and expenses	Servicing Fee: 35 bps per annum; and Extraordinary fee: up to 25 bps per annum.
Delinquency stress	Delinquency Level: Two-third of the WAFF in each of the first 36 months; and Recovery Timing: 36 months.

Table 19

Default Timing Curves (% Of WAFF)
Year number*	Front-loaded (%)	Back-loaded (%)
1	5	5
2	15	5
3	20	10
4	25	15
5	15	20
6	10	15
7	5	15
8	5	10
9		5
Total	100	100
*Percentage of weighted-average foreclosure frequency applied in each term. For modeling purposes, the percentage in each stage divided by 12.

Table 20

Prepayment Rates (CPR)
	Low	High
CPR(%)	3	15
CPR--Constant prepayment rate.

Details Of Further Considerations When Applying The Cash Flow Criteria To The Analysis Of A Chinese RMBS Transaction

In our rating analysis, we also assess a transaction's payment structure and cash flow mechanics. This uses our own quantitative models to evaluate whether the cash flows from the assets suffice, at the applicable rating levels, for making timely payments of interest and ultimate payment of principal (i.e., by or before the legal maturity date).

In our cash flow analysis, we use the pool-level WAFF and WALS described in previous sections to reflect credit stress at each rating level.

During modeling, cash flow stresses test the credit and liquidity support the assets need to make timely interest payments and final principal payments by their legal maturity dates. Our models consider any available structural support, such as cash reserves, liquidity facilities, and hedging arrangements.

For revolving stand-alone RMBS structures (i.e. structures backed by a pool whose assets change or revolve), the modeling approach aims to reflect the structure after the activation of any " early amortization" (or stop-substitution) triggers and applies cash flow stresses from this point. An early amortization trigger is an event or situation that halts the substitution of assets in a revolving loan pool.

In our surveillance of existing ratings, cash flow modeling may show that under stress at the 'B' rating level, a particular tranche will miss interest payments or fail to repay the principal by or before the final legal maturity date. If this is the case, then our initial assessment, assuming all factors remain the same, may be to consider lowering the rating on those securities to 'B-' or lower.

Depending on our view of a transaction's immediate cash flow position, the rating could move into the 'CCC', 'CC', or 'C' rating category, consistent with our ratings definitions, and in accordance with "Criteria For Assigning 'CCC+', 'CCC', 'CCC-', And 'CC' Ratings," Oct. 1, 2012.

Default timing

The cumulative amount of defaults for cash flow modeling is equal to the pool's WAFF, whereas the cumulative amount of recoveries is calculated as (1 minus WALS).

For certain structure types (for example, master trust transactions), we may run additional analysis to test the sensitivity of the liability structures to different default timing curves.

In some cases, we may make an adjustment to the default timing curves to better reflect the pool's composition and transaction structure. For example, we may shorten the default curves where the mortgage portfolio's average remaining term gets shorter.

Foreclosure period assumptions represent the estimated time to repossess and sell a property upon a default (see table 18 above). They also reflect the typical time necessary for judicial proceedings, delivery of forward-delivery houses, and any other likely delay.

Cash flow modeling considers the negative carry resulting from interest due on the rated liabilities during the foreclosure period.

The loss-severity estimates used in the cash flow modeling are based on the loan principal and assume no recovery of interest accrued on the mortgage loans during the foreclosure period.

Interest rate scenarios

Specific structural features may involve using additional cash flow stresses, such as alternative interest rate patterns or different default-timing curves, among others.

Prepayment scenarios

Residential loan prepayments vary the amount of excess spread available and this may affect the absolute level of defaults exhibited in a transaction.

When analyzing the payment structure and cash flow mechanics of Chinese RMBS, we typically test the transaction's ability to withstand high and low prepayment scenarios as set out in table 20.

We may raise the prepayment assumptions if a pool's historical prepayment rates were higher than historical averages or if a transaction were particularly sensitive to prepayment risk. We may also reduce prepayment stress in situations where long-term historical data support lower prepayment rate assumptions for a specific loan product.

Modeling of reinvestment rates and spread compression

We typically do not assume any revenues associated with any cash or securities the issuer holds, unless there is an eligible guaranteed investment contract (GIC) in place.

On a case-by-case basis, we apply a spread compression assumption if we identify a risk of material decrease in spread over time. For example, borrowers paying higher interest rates might tend to prepay earlier or indicate higher default risk.

Originator insolvency, commingling, and set-off

The analysis of any commingling or setoff risks that can result from an originator's or servicer's insolvency follows the application of the counterparty criteria (see "Counterparty Risk Framework: Methodology And Assumptions," March 8, 2019). The counterparty criteria determine the maximum supported rating based on available mitigation of such counterparty risks. Mitigants may include a counterparty's commitment to implement remedies upon a downgrade, or structural factors such as the coverage of the risk through credit enhancement, demonstrated through the modeling of any exposure.

Regarding commingling risk, the degree to which a collection account holder's insolvency affects the cash flow from the assets in a mortgage loan pool depends on the collection account's characteristics, if the collection account is not in the name of the RMBS issuer.

The amount at risk depends on the timing of scheduled payments from borrowers, the frequency of transfers into the transaction account, and the level of prepayments.

Modeling of senior fees and expenses

The modeling of all an issuer's foreseeable expenses uses stressed costs to reflect the need to replace the initial service provider.

The most significant portion of senior fees and expenses is related to the servicer. Estimation of transaction maintenance fees usually includes a certain increase from the amount stipulated in the transaction agreement to account for a potential replacement of the servicer. We also examine the degree to which the structure is typical, and the levels of servicing fees in the market. The estimate of fee assumptions in table 18 reflects our findings.

To reflect the likely cost of replacing the initial servicer, we set servicing fees at the higher of the contractual rate and the assumed rates in the table 18.

Other unexpected expenses may arise during the life of a rated security, such as when amending transaction documents, costs associated with perfection of title, and any other unexpected costs required for the orderly maintenance of the assets. In the absence of any structural mechanism, such as a reserve designated to cover substantial or full amount of the unexpected expenses, an expense cap, or if the cap is significantly lower than industry standard, we might assume extraordinary expenses of up to a certain amount in our cash flow analysis (see table 18 above).

We may apply different assumptions from those detailed in table 18. For example, we may lower the level to reflect easier/simpler tasks of the trustee/servicer. On the other hand, we may raise the level for the exotic transaction to reflect a difficulty of the trustee/servicer's replacement.

Delinquency

We assume a delay of a proportion of scheduled interest and principal receipts in each of the first several months of a hypothetical recession, and set full recovery of the arrears to take place several months after the delinquency occurs (see table 18 above). The cash flow stress for delinquencies is independent of the arrears adjustment to the WAFF.

For pools that contain residential loans with an option to temporarily suspend periodic payments (payment holiday loans), we assume a delay of a proportion of scheduled interest and principal receipts. In situations where there is the potential for payment holidays to be granted after a loan's inception or where payment holidays have been granted (e.g., due to governments' and banks' forbearance measures for households and small and midsize corporates), we may apply an additional stress in our cash flow analysis where relevant. In those instances, we base our assumption for the delay to scheduled principal and interest receipts on an estimate of the proportion of a pool that opts to take a payment holiday and the likely duration of the holiday. The likely duration will be assessed with reference to factors that may include, but are not limited to, relevant legislative frameworks, collateral credit quality, servicers' policies, and available servicer data on payment holidays granted.

CRITERIA SUPPLEMENT: JAPAN

Variables Used As An Input In Our Calculation Of The Foreclosure Frequency

Japan mortgage market assessment (MMA)

Our MMA in a given jurisdiction consists of an assessment on a six-point scale ranging from "very low risk" to "extremely high risk".

Table 21

Mortgage market assessment
Japan MMA	Low risk
Sub-factors
Economic risk	Low risk
Mortgage industry risk	Intermediate risk
Lender - recourse assessment	Full recourse

As per our MMA methodology framework (see Appendix I and table 5 of our residential mortgage criteria), the MMA for Japan reflects our view of:

• The risks in the Japanese banking system;
• Japan's low unemployment sensitivity to changes in economic output, combined with a sophisticated social welfare system; and
• Borrowers' willingness to pay in low/negative equity scenarios because of the full-recourse associated with most Japanese mortgage products.

The Japan archetypal foreclosure frequency anchors

When applying paragraphs 15 and 26, and table 1 of our residential mortgage criteria, based on an MMA of "low risk" ('2'), we set Japan's foreclosure frequency assumptions as indicated in table 22 below. As per paragraphs 23-25 of our residential mortgage criteria, the 'B' foreclosure frequency assumption reflects our assessment of historical performance and our expectations for future performance over the medium term given anticipated macroeconomic conditions. In our analysis, we considered default expectations across mortgage portfolios and the performance of outstanding Japanese RMBS transactions we rate, in both cases adjusted for seasoning to reflect expected lifetime defaults.

Table 22

Archetypal foreclosure frequency anchors
Rating level*	Archetypal foreclosure frequency (%)
AAA	10.0
AA	7.5
A	5.0
BBB	3.2
BB	2.1
B	1.1
*Assumptions for intermediate rating levels are interpolated.

The archetypal Japanese pool

We define the archetypal pool for Japan as follows:

Table 23

Japan--archetypal pool
Characteristics by type	Archetypal features
Pool
Pool size	At least 250 loans at issuance for owner-occupied or condominium investment loans. At least 40 effective loan counts for apartment loans at issuance.
Originator	No adjustment factor related to the quality of the lender's underwriting or for features not specifically covered by other adjustments.
Geographic distribution	Diversified nationally.
Borrower
Borrower type	Borrower is a private individual and is not self-employed.
Citizenship	Japanese citizen or a person who has status of permanent resident in Japan.
Performance status	Not delinquent.
Borrower credit history	No adverse credit history.
Affordability	Lender has assessed the borrower's income.
Loan
Seasoning	Up to 60 months
Loan amortization profile	Fully amortizing
Loan product	No payment shock feature
Loan purpose	Purchase or to refinance the balance on an existing loan (where the lender has fully re-underwritten the loan) of a residential property for owner occupation.
Loan-to-value (LTV)	85% (calculated weighting the original LTV and current indexed LTV in an 80:20 ratio).
Debt-to-income (DTI)	25%
Security	First-lien mortgage on the property.
Interest rate	Fixed rate
Group credit life insurance	Insured for owner-occupied or condominium investment loans. Not insured for apartment loans.
Property
Property type	Residential
Occupancy status	Owner occupied and primary residence.
Valuation method	Purchase price or lender's valuation.
Valuation amount	Up to the applicable jumbo valuation threshold specified in table 28.

Variables Considered When Calculating The Foreclosure Frequency Of A Mortgage Pool

For each characteristic and attribute that differ from the archetype at a pool, loan, or borrower level, the criteria apply a corresponding adjustment to the foreclosure frequency. The following table shows the applicable variables.

Table 24

Variables considered when calculating the foreclosure frequency of a mortgage pool
Loan-to-value (LTV)	Type 3 LTV curve (full-recourse market; see the chart titled "LTV Curve " below).
Combined LTV definition	The LTV is calculated weighting 80% of the original LTV (OLTV) and 20% of the current LTV (CLTV).
Originator adjustment	Typically, 0.7x-1.3x or higher applied at pool level. For country specific originator adjustment factors see "Originator adjustment" in "Details Of Further Considerations When Calculating The Foreclosure Frequency Of A Japanese Mortgage Pool" below.
Loan affordability:
Debt-to-income (DTI)	0.8x if DTI <= 15%; 1.0x if DTI = 25%; 1.2x if DTI = 35%; 3.0x if DTI => 53%. This adjustment is a continuous function (see the chart titled "DTI Curve" below).
Borrower occupancy status: Owner-occupied (second home)	1.3x
Borrower occupancy status: Condominium investment loan	1.5x. In addition, we apply adjustments of 1.0x–1.5x based on the borrower's income to reflect the increased risk of borrowers with lower income levels who may engage in more speculative property investments (see table 25). If no adequate data on a borrower's income exist, an alternative adjustment factor of 1.5x is applied to such a loan. This is equal to the highest adjustment factor in the table.
Borrower occupancy status: Apartment loan	3.0x – 6.0x. If no adequate DSCR data are available, an alternative adjustment factor of 6.0x applies for apartment loans, which equals the highest adjustment factor. See the chart titled "Adjustments For Apartment Loans By Debt Service Coverage Ratio" below.
Employment status	See table 26. If a borrower's self-certified information on regular or nonregular employment status is not available, we apply a 1.1x adjustment to general (full-time) salaried employees' mortgage loans. If no adequate data on the tenure of self-employed borrowers are available, we apply an alternative adjustment factor of 2.0x for such a loan, which is equal to the adjustment factor in the table for a borrower who is self-employed for one to two years.
Payment shock (Loan product type)	1.0x for full-term fixed rate; 1.1x for full-term fixed rate with step-up if monthly payment amount increases to 125% or less of previous amount. Applied until 12 months after the final step-up; 1.2x for full-term fixed rate with step-up if monthly payment amount increases to more than 125% of previous amount. Applied until 12 months after the final step-up; 1.2x for full-term floating rate; 1.1x for convertible with a 10-year or more initial fixed-rate period; 1.2x for convertible with a less than 10-year initial fixed-rate period. Convertible refers to fixed/floating rate convertible loans that allow borrowers to choose a fixed-interest rate period or a floating rate after the fixed-interest term expires.
Geographic concentration	1.2x, which is applied to the exposure in excess of the regional concentration limits outlined in table 27.
Second lien loans	1.0x where second-lien loans are extended from lenders that also hold the associated first-lien loans or first-lien loans are provided by the public financial sector/employee loan program. 1.5x where there is sufficient data available to back up the other lien loans; 1.7x where there is insufficient data available to back up the other lien loans.
Loan purpose	1.2x for refinancing that involves features such as home equity, a cash take-out, or a loan extension.
Arrears	1.0x for loans that are current or currently one month delinquent; 2.5x for loans currently two months delinquent; 5.0x for loans currently three months delinquent; 100% foreclosure frequency for loans currently 4 months or more delinquent.
Seasoning (adjustment factors for loan seasoning)	1.0x for seasoning <=5 years; 0.75x for seasoning >5 and <=6 years; 0.70x for seasoning >6 and <=7 years; 0.65x for seasoning >7 and <=8 years; 0.60x for seasoning >8 and <=9 years; 0.55x for seasoning >9 and <=10 years; 0.50x for seasoning >10 years. The above factors apply only to loans that are current or in arrears for up to one month.
Group credit life insurance adjustment	1.05x if <= 35 years old; 1.1x if > 35 to < 46 years old; 1.3x if >= 46 years old. The above factors apply based on weighted-average borrower age of the pool at closing.

Table 25

Adjustment for condominium investment loans
Borrower's annual income	Adjustment factor (based on income; x)
<= ¥5 million	1.5
>¥5 million to <=¥6 million	1.25
>¥6 million to <=¥7 million	1.1
>¥7 million	1.0

Table 26

Adjustment for employment status
Employment status	Adjustment factor (x)
Regular salaried, licensee	1.0
Pension or over 65 at loan origination	1.5
Commission-based	2.0
Nonregular salaried (full-time)	2.0
Nonregular salaried (part-time)	3.0
Unemployed, other	4.0
Self-employed for:
<= one year	3.0
> one year <= two years	2.0
> two years <= three years	1.5
> three years <= four years	1.2
> four years <= five years	1.2
> five years	1.0

Table 27

Mortgage loan concentration limits by region
Region	Concentration limit (%)
Hokkaido	8
Tohoku	14
Kanto	69
Chubu	34
Kinki	35
Chugoku	12
Shikoku	6
Kyushu	23

Details Of Further Considerations When Calculating The Foreclosure Frequency Of A Japanese Mortgage Pool

Originator adjustment

Specific examples of the factors considered in determining the originator adjustment for a pool of Japanese residential loans typically include but are not limited to:

• Transactions without updated loan-by-loan (LBL) data available during surveillance until the weighted average seasoning of loans reaches five years;
• Borrowers in arrears in the case of transactions without updated LBL data during surveillance;
• Apartment loan borrowers who are insured by group credit life insurance;
• Borrowers with no income verification;
• Loans with lower documentation standards;
• Loans with a full valuation (instead of a lender's valuation);
• Interest only (I/O) or balloon payment loans;
• Selection bias to reperforming and restructured/modified loans;
• Owner-occupied loans financing mixed properties with more than 50% commercial occupation;
• Commercial borrowers in owner-occupied or condominium investment loans;
• Borrowers with adverse credit histories;
• Loans with an initial term to maturity that is longer than the market standard;
• Borrowers who are not Japanese citizens or do not have permanent resident status in Japan;
• Any other observed risk characteristics that are not adjusted for at a loan-by-loan level;
• Significant exposure to single obligors; and
• Changing in underwriting guidelines for which impact is not yet visible in performance data.

Loan to value (LTV)

LTV is calculated in a three-stage process. Stage 1: The original LTV (OLTV) is calculated using the original loan balance and the property valuation at the time of loan origination. Stage 2: The current LTV (CLTV) is calculated using the loan balance as of the portfolio cut-off date and the current indexed property value. Stage 3: The LTV is calculated by using the specific weighting of the OLTV and CLTV as described in table 24. The OLTV and CLTV are based on the full loan balances secured on the property, including loan parts outside the asset pool and prior- and second-ranking balances, if any.

A loan's CLTV is the current balance of a loan, reflecting the actual principal pay down, divided by the indexed initial value or other updated value of the property. For transactions without updated loan-by-loan (LBL) data available during surveillance, we use a hypothetical current loan balance reflecting the scheduled amortization to calculate the CLTV.

Loan affordability: Debt-to-income (DTI)

In calculating the ratio of DTI, we incorporate any secondary income such as rental income from investment properties or a guarantor's income where appropriate.

For apartment loan RMBS, the primary source of repayment is the rental income from the apartment properties, rather than the borrower's salaried income. Therefore, a debt service coverage ratio (DSCR) adjustment is applied in place of the DTI adjustment.

Borrower occupancy status (investment properties)

The investment property classification includes second homes, investments in condominium units for lease, and investments in apartment buildings for lease.

A typical borrower of an apartment loan is a landowner who builds a rental apartment building for investment and/or inheritance tax purposes. Unlike owner-occupied prime mortgage loans in Japan, most apartment loans in the country are extended on a nonrecourse basis.

We apply an adjustment factor based on an apartment loan's actual DSCR (see the chart 6).

The borrower's capacity to pay for apartment loans relies more on rental income from the property than earnings from his or her job. Thus, we do not apply adjustment factors based on DTI, employment status, or loan seasoning (as these are described above or below) for Japanese RMBS backed by apartment loans.

Employment status

We assume all loans in the Japanese mortgage market are income verified. If the pool includes loans with no income verification, an additional risk will be captured in the originator adjustment.

For self-employed borrowers, we apply an adjustment based on the tenure of self-employed borrowers, subject to availability of such data.

For apartment loan RMBS, the major source of repayment is a rental income from the apartment properties, not a borrower's stable salary income. Therefore, adjustment for employment status is not applicable.

Geographic concentration

If a pool has significant geographic concentration risk that we believe is not sufficiently captured, we may capture it using the originator adjustment.

Seasoning (adjustment factors for loan seasoning)

Adjustments for loan seasoning are provided in table 24. The seasoned term of a refinancing loan includes the years elapsed on the previous (original) loan (i.e., the seasoned period prior to refinancing).

For transactions without updated LBL data during surveillance, we do not apply a seasoning adjustment factor. Not doing so mitigates risk arising from the lack of information about loans that are still outstanding in the pool, as well as the current outstanding amount of each loan.

For apartment loan RMBS, the primary source of repayment is not a borrower's stable income, but rental income from the apartment properties. Repayment history does not necessarily mean an improvement in credit quality, in our view. Therefore, we do not apply the seasoning adjustment factor.

Loans in arrears

When only pool-level performance data are available and no adequate data are available about which loans are delinquent during surveillance, this adjustment is not applied. Instead, the risk related to pool-level actual arrears is captured in the originator adjustment.

Payment shock (loan product type)

Adjustments are applied to loans considered to have a potential for an increase in monthly payments. Loans where those principal payments start within a year after loan origination are not treated as loans with payment shock or interest only (IO) loans, because this is one of the standard mortgage products in Japan.

Second-lien loans

We apply an adjustment factor for second-lien mortgage loans as indicated in table 24. We differentiate a variable by the other lien loan's data availability.

Borrowers in Japan who are eligible for mortgages can borrow 100% of the mortgaged property's purchase price in general, so second-lien loans are not common in the Japanese market. Borrowers who want to borrow two loans, one with a fixed interest rate and the other floating, or a part of their required funds from the public financial sector/employee loan program at a lower interest rate use second-lien loans. In those cases, we believe the foreclosure frequencies of such borrowers do not differ from borrowers who purchase houses using first-lien loans only.

Small pool adjustment for apartment loan pools

In the case of apartment loan RMBS, we apply a small-pool adjustment for apartment loan pools in this section rather than a small pool adjustment described above in our residential mortgage criteria.

We define a sufficiently diversified Japanese apartment loan RMBS as one that has an effective loan count of 40 or more at a transaction's closing. We analyze and measure the effective loan count using the Herfindahl Index (see chart below) to assess the diversity of apartment loan RMBS. Specifically, the inverse of the Herfindahl Index is often used to count the effective number of loans over which the pool balance is distributed. We do not assign ratings to apartment loan RMBS with effective loan counts of less than 40 at the transaction's closing under this methodology.

Group credit life insurance adjustment

We apply adjustments for group credit life insurance if: i) borrowers do not take out group credit life insurance, ii) borrowers have a right to terminate it at their sole discretion, or iii) a transaction agreement stipulates that the transaction party terminates the group credit life insurance in the event of the originator's insolvency.

We may make an adjustment to the variables shown in this document to better reflect the pool's composition and transaction characteristics. For example, we may lower the variable in case of high weighted average foreclosure frequency (WAFF) for a pool as well as in case of mixed pools where one portion of the loans are from borrowers who have taken out life insurance, and one portion is not. Conversely, we may raise the variable in the case of a low pool WAFF.

Traditional apartment loans are not insured by group credit life insurance because of their nonrecourse nature. We therefore do not apply this adjustment to apartment loans.

Variables Considered When Calculating The Loss Severity Of A Japanese Mortgage Pool

The table below shows the variables we consider when calculating the loss severity of a mortgage pool.

Table 28

Variables considered when calculating the loss severity of a Japanese mortgage pool
Property indexation	To estimate price fluctuations for each property since loan origination, we use the Japanese government's official land market value publication ("Koji Chika"), published every March since 1970, to provide basic information about land prices in Japan. We base property value assumptions on prefecture-specific index data for residential areas. We reflect changes in the index without adjustment (i.e., we reflect 100% of the changes in the index, irrespective of the direction).
Jumbo valuation threshold*	¥100 million
Valuation haircuts	Typically, 0%
Foreclosure costs¶	¥1 million fixed cost. 5% variable as a percentage of the post-repo, jumbo-adjustment valuation.
Foreclosure timeline/period	18 months for residential properties (primary and secondary homes) and condominiums; 24 months for apartment properties.
*We increase MVD assumptions for jumbo valuations. We apply an adjustment of 20% on the excess above the jumbo threshold. For example, for a property with an indexed-valuation of ¥150 million, we apply an adjustment on the difference between ¥150 million and ¥100 million (jumbo valuation threshold). Jumbo valuation adjustment, the product of 20% x ¥50 million in this example, is then deducted from the indexed valuation after applying the post-repossession market value decline. ¶The post-repo, jumbo-adjustment valuation is the property value obtained after the application of the repossession market value decline at the relevant rating as well as the jumbo valuation adjustment to the indexed-valuation.

Details Of Further Considerations When Calculating The Loss Severity Of A Japanese Mortgage Pool

Forced-sale discount (FSD)

On a case-by-case basis, we may increase or decrease the forced sales discount where there is sufficient information to support it being either higher or lower than envisaged using the standard calculation.

Current loan to value

For the purpose of the weighted average loss severity (WALS), a loan's CLTV is the current balance of a loan, reflecting the actual principal pay down, divided by the indexed initial value or other updated value of the property.

For transactions without updated LBL during surveillance, we use the hypothetical current loan balance, reflecting the scheduled amortization, to calculate a loan's CLTV.

Property valuations

For owner-occupied residential loans and condominium investment loans, the valuation of the collateral property uses available data in the order below. As the level of down payments relative to the purchase price is a more relevant driver to assess the default risk of a borrower, we give priority to the purchase price over the lender's valuation.

• Full appraisal
• Purchase price
• Lender's valuation

For apartment loans, the valuation of the collateral property uses available data in the order below. This is because investors usually refer both values indicated by the income approach and the cost approach.

• Full appraisal
• Whichever is lower of the purchase price or lender's valuation

Valuation haircut

A standard valuation method in Japan is a purchase price or a lender's valuation because a full appraisal is not common in the Japanese mortgage market. Given this, we generally do not apply a haircut based on the valuation type. Where relevant, we may make an adjustment to better reflect the pool's characteristics. For example, we may revise up or down a lender's valuation if they employ less or more conservative valuation methods relative to market standards.

Full appraisals are generally more conservative than purchase prices or lenders' valuations. If the pool includes loans with full appraisals, the correspondingly more conservative underwriting compared with market standards will be captured in the originator adjustment.

Mortgage guarantees

Some residential loans in Japan are secured by a mortgage guarantor or insurer. In such cases, we apply the methodology outlined in "Methodology For Assessing Mortgage Insurance And Similar Guarantees And Supports In Structured And Public Sector Finance And Covered Bonds," Dec. 7, 2014.

In cases where a mortgage guarantor or insurer does not have an issuer credit rating, we take into consideration the possibility of recoveries from collateral properties if the guarantor creates a mortgage right to secure its recourse rights against the obligor.

In cases where a transaction structure has been established to transfer the amount recovered from the mortgage, we will review the effectiveness of the structure. If the structure effectively allows recoveries from the mortgage to be transferred to creditors, we incorporate the entire estimated amount to be collected from the collateral properties into our analysis. Where such a structure has not been established, we incorporate half of the estimated amount to be collected from the collateral properties into our analysis in cases where the parent company of a guarantor is a regulated financial institution. In all other cases, we do not incorporate recoveries.

Details Of Further Considerations When Calculating Minimum Credit Enhancement Levels In Japanese RMBS Or Covered Bond Transactions

Pro-rata pay structure and tail risk in JHF RMBS

The government-related entity (GRE) Japan Housing Finance Agency's (JHF) RMBS employs a pro rata pay structure from inception without subordination floors. To mitigate unexpected event risk in the tail end, in addition to the credit enhancement floors outlined in "credit enhancement and subordination floors" section of our residential mortgage criteria, we apply in our analysis:

• A 1.2x multiple to the foreclosure frequency assumptions during pro rata; and
• A minimum loss assumption at each rating category, based on a certain number of loans defaulting and liquidating as outlined in the table below.

Table 29

Rating category minimum loss assumptions for transactions with fewer than 100 loans
Maximum potential rating	Minimum loss assumptions
AAA	Highest-balance loan liquidated at the greater of 50% loss severity and the applicable 'AAA' loss severity, plus the higher of the next nine largest loss exposures at the 'AAA' loss severity and 20%.
AA category	Highest-balance loan liquidated at the greater of 50% loss severity and the applicable 'AA' loss severity, plus the higher of the next seven largest loss exposures at the 'AA' loss severity and 20%.
A category	Highest-balance loan liquidated at the greater of 50% loss severity and the applicable 'A' loss severity, plus the higher of the next five largest loss exposures at the 'A' loss severity and 20%.
BBB category	Highest-balance loan liquidated at the greater of 50% loss severity and the applicable 'BBB' loss severity, plus the higher of the next three largest loss exposures at the 'BBB' loss severity and 20%.
BB category	Highest-balance loan liquidated at the greater of 50% loss severity and the applicable 'BB' loss severity, plus the higher of the next largest loss exposure at the 'BB' loss severity and 20%.
B category	Highest-balance loan liquidated at the greater of 50% loss severity and the applicable 'B' loss severity.

We view the unique structure of JHF RMBS (including the removal of default loans, four-month delinquent loans, and loan-term modified loans prior to a beneficiary certificate trigger event, a trigger to switch to sequential from pro rata upon an occurrence of trigger events), along with the application of the additional minimum loss assumptions and WAFF adjustments described herein, as providing sufficient mitigants as described in "credit enhancement and subordination floors" section of our residential mortgage criteria. Therefore, subordination floors described in table 4 of our residential mortgage criteria are not applicable to JHF RMBS.

Pro rata pay structure and tail risk in non-JHF RMBS

In non-JHF RMBS, the subordination floors described in "credit enhancement and subordination floors" section of our residential mortgage criteria generally mitigate tail risk associated with pro rata pay transactions from inception. However, to compensate for any unexpected risks, we would in addition apply a 1.2x multiple to the foreclosure frequency assumptions in our cash flow analysis for transactions without updated LBL data during surveillance that are pro rata or may become so in the future.

Additional credit enhancement floors for apartment loan RMBS

Apartment loan RMBS sometimes contain very large loans in the underlying pools. To ensure rating stability in case a larger loan defaults, we establish a credit enhancement floor for apartment loan RMBS as shown in the table below. Transactions in higher rating categories should withstand a certain number of large-loan defaults in the portfolio. Based on this table, credit enhancement for 'AAA' rated RMBS backed by apartment loans must be larger than the aggregate balance of the five largest loans. After taking into account loss severity (i.e., recovery assumptions from the defaulted loans under a 'AAA' stress scenario), the 'AAA' rated securities should withstand a net loss of the 10 largest loans. We can consider soft credit enhancement, such as excess spread, if appropriate.

Table 30

Number of largest apartment loans with default exposure by rating category
	AAA	AA category	A Category	BBB Category	BB Category	B Category
Gross loss	Top 5	Top 4	Top 3	Top 2	Top 1	N/A
Net loss	Top 10	Top 8	Top 6	Top 4	Top 3	Top 2
N/A--Not applicable.

In some cases, we may adjust gross loss or net loss thresholds for each rating category by changing the number of loans to be covered. For example, we may lower a threshold when the remaining life of the transaction is short or in cases where the threshold breach would in our view quickly resolve.

Variables Considered In The Cash Flow Model Runs

Table 31

Variables considered in the cash flow model runs for a Japanese mortgage pool
Default timing scenarios	Default timing curves: Typically, "front-loaded" and "back-loaded". In both scenarios, the defaults occur over a 20-year period, with each scenario having a different peak timing for defaults. The peak level of stress in each scenario is referred to as a recession. It corresponds to months 61-120 (in the case of the front-loaded scenario) and months 121-180 (in the case of the back-loaded scenario). See table 32.
Interest rate scenarios	Standard run scenarios: Typically, "up", "up-down" and "down" (See "Methodology To Derive Stressed Interest Rates In Structured Finance," Oct. 18, 2019. These curves vary by stress scenario).
Basis risk	Basis risk stress timing: We apply the stress corresponding to a rating level in the cash flow analysis for the first 18 months. After this 18-month period, the 'B' percentile applies at all rating levels.
Prepayment scenarios	Typically, "Low" and "High". Prepayment scenarios generally start at 3.0% per year at transaction closing and increase on a linear basis for up to five years. See table 33.
Senior fees and expenses	See table 34.
Additional costs for group credit life insurance	30 basis points if <= 35 years old; 45 basis points if > 35 to < 46 years old; 100 basis points if >= 46 years old. The additional costs apply based on weighted-average borrower age of the pool at closing.

Table 32

Default timing curves (% Of WAFF)
Number of months*	Front-loaded (%)	Back-loaded (%)
1-60	35	5
61-120	45	40
121-180	15	40
181-240	5	15
Total	100	100
*Percentage of weighted-average foreclosure frequency applied in each term. For modeling purposes, the percentage in each stage divided by 60 is applied each month. WAFF--Weighted average foreclosure frequency.

Table 33

Prepayment rates
Annual (%)	Low	High
Start	3	3
Year 5 and thereafter	3	12

Table 34

Senior Fees And Expenses
Category	Fee (bps)*
Trustee fee
Initial loan pool outstanding exceeds or equals ¥30 billion	5
Initial loan pool outstanding is below ¥30 billion	10
Servicing fee
Prime-lender origination (e.g., major banks, trust banks, regional banks, or labor banks)	35
Appointed backup servicer at closing	35
No appointed backup servicer at closing and initial loan pool outstanding exceeds or equals ¥30 billion	40
No appointed backup servicer at closing and initial loan pool outstanding is below ¥30 billion	45
*We use the contractual rate if it is higher than the rate in the table. bps--Basis points.

Details Of Further Considerations When Applying Our Cash Flow Criteria To Analysis Of Japanese RMBS Or Covered Bond Transactions

In our rating analysis, we also analyze a transaction's payment structure and cash flow mechanics. This analysis uses our own quantitative models to assess whether the cash flows from the assets suffice, at the applicable rating levels, for making timely payments of interest and ultimate payment of principal (i.e., by or before the legal maturity date).

In our cash flow analysis, we use as inputs the pool-level WAFF and WALS as described in our residential mortgage criteria and previous sections above, to reflect credit stress at each rating level.

During modeling, cash flow stresses test the credit and liquidity support the assets need. Our models consider any available structural support, such as cash reserves, liquidity facilities, and hedging arrangements.

In our surveillance of existing ratings, cash flow modeling may show that under the 'B' stress a particular tranche will miss interest payments or fail to repay the principal by or before the final legal maturity date. If this is the case, then our initial assessment, assuming all factors remaining the same, may be to consider lowering the rating on those securities to 'B-' or lower.

Depending on our view of a transaction's immediate cash flow position, the rating could move into the 'CCC', 'CC', or 'C' rating category, consistent with our ratings definitions, and in accordance with "Criteria For Assigning 'CCC+', 'CCC', 'CCC-', And 'CC' Ratings," Oct. 1, 2012.

Default timing

The cumulative amount of defaults for cash flow modeling is equal to the pool's WAFF, whereas the cumulative amount of recoveries is calculated as (1 – WALS).

For certain structure types (for example, master trust transactions), we may run additional analysis to test the sensitivity of the liability structures to different default timing curves. In some cases, we may make an adjustment to the default timing curves to better reflect the pool's composition and transaction structure. For example, we may shorten the default curves where the mortgage portfolio's average remaining term gets shorter.

Foreclosure period assumptions represent the estimated time to repossess and sell a property upon a default. They also reflect the typical time necessary for judicial proceedings and any other likely delay.

Cash flow modeling considers the negative carry resulting from interest due on the rated liabilities during the foreclosure period.

The loss severity estimates used in the cash flow modeling are based on the loan principal and assume no recovery of interest accrued on the mortgage loans during the foreclosure period.

Interest rate scenarios

If a transaction features an interest rate cap--either on the note coupon or through a cap agreement with an external counterparty--upward interest rate stress assumptions exceeding the cap level may be unduly beneficial for the transaction's cash flow projection. In such events, we may apply a different interest rate stress to test sensitivity of the ratings in the absence of the cap.

Specific structural features may involve using additional cash flow stresses, such as alternative interest rate patterns or different default timing curves. For example, if we consider a transaction to have unique characteristics that lead to different sensitivity to interest rate scenarios, we may apply an additional scenario as well.

Basis risk

We adjust the interest spread modeled between assets and liabilities by applying different spreads over the life of a transaction as a haircut to the margin. The size of the spreads typically depends on the distribution of historical differences among indices, using the rating-specific values corresponding to the percentiles shown in the table below.

Table 35

Basis risk percentile stresses
Rating category	Percentile (%)
AAA	95
AA	90
A	65
BBB	50
BB	40
B	30

Prepayment scenarios

Residential loan prepayments vary the amount of excess spread available and this may affect the absolute level of defaults exhibited in a transaction.

When analyzing the payment structure and cash flow mechanics of Japanese RMBS, we typically test the transaction's ability to withstand high and low prepayment scenarios as set out in table 33.

We may raise the prepayment assumptions if a pool's historical prepayment rates were higher than historical averages or if a transaction were particularly sensitive to prepayment risk (e.g., excess spread notes). We may also reduce prepayment stress in situations where long-term historical data support lower prepayment rate assumptions for a specific loan product.

Modeling of reinvestment rates and spread compression

We typically do not assume any revenues associated with any cash or securities the issuer holds. This is because a general practice in Japan is for issuers to hold any cash on the noninterest bearing account.

Also, we typically do not assume any spread compressions on the loan pool in our cash flow analysis. This is because risk-based pricing is not common in standard mortgage underwriting practice in Japan. As a consequence, all loan spreads are very similar in each pool.

Originator insolvency, commingling, and set-off

The analysis of any commingling or setoff risks that can result from an originator's or servicer's insolvency follows the application of the counterparty criteria (see "Counterparty Risk Framework: Methodology And Assumptions," March 8, 2019). The counterparty criteria determine the maximum supported rating based on available mitigation of such counterparty risks. Mitigants may include a counterparty's commitment to implement remedies upon a downgrade, or structural factors such as the coverage of the risk through credit enhancement, demonstrated through the modeling of any exposure. In the context of Japanese RMBS transactions, government-sponsored deposit insurance also often mitigates commingling and setoff risks. If we find that deposit insurance fully mitigates the risk, the rating will not be constrained under counterparty criteria. Where our analysis relies on government-sponsored deposit insurance to cover commingling or setoff risks, we analyze the resulting sensitivity to sovereign risk under our criteria for rating above the sovereign (see "Incorporating Sovereign Risk In Rating Structured Finance Securities: Methodology And Assumptions," Jan. 30, 2019).

Regarding commingling risk, the degree to which a collection account holder's insolvency affects the cash flow from the assets in a mortgage loan pool depends on the collection account's characteristics, if the collection account is not in the name of the RMBS issuer.

The amount at risk depends on the timing of scheduled payments from borrowers, the frequency of transfers into the transaction account, and the level of prepayments.

If a Japanese RMBS transaction satisfies the below five conditions, we assume that commingling risk at an originator deposit-taking institution in Japan is mitigated by the Deposit Insurance Corp. of Japan (DICJ), a government-sponsored insurer:

• The servicer falls under the jurisdiction of the deposit insurance law in Japan;
• The loan agreement specifies that loan transfers are allowed (borrower's recognition);
• The transaction scheme specifies that the deposit-taking institution entrusts the loans with the special-purpose company and continues loan servicing operations as a servicer as dictated by an administrative agreement;
• The servicer appropriately manages proceeds by separating the collection of securitized loans from its own funds; and
• The payments are denominated in Japanese yen.

For an originator that is a deposit-taking institution in Japan, we consider setoff risk to be mitigated by the DICJ up to the maximum insured amounts of the borrower's deposit amounts.

Modeling of senior fees and expenses

The modeling of all an issuer's foreseeable expenses uses stressed costs to reflect the need to replace the initial service provider.

The most significant portion of senior fees and expenses is related to the trustee, the servicer, and the backup servicer, if involved. Estimation of transaction maintenance fees usually includes a certain increase from the amount stipulated in the transaction agreement to account for a potential replacement of the trustee or servicer. We also examine the size of the transaction, the degree to which the structure is typical, and the levels of trustee fees in the market. The estimate of maintenance fees reflects our findings.

To reflect the likely cost of replacing the initial servicer, we set servicing fees at the higher of the contractual rate and the estimated rates depending on the size of the transaction and the appointment of a backup servicer as in table 34.

We may apply different assumptions from those detailed in table 34. For example, we may lower the level for a repackaged transaction to reflect easier/simpler tasks of the trustee/servicer. On the other hand, we may raise the level for the exotic transaction to reflect a difficulty of the trustee/servicer's replacement.

Additional costs for group credit life insurance

For transactions in which group credit life insurance premium payments are included in the waterfall, the assumption incorporates an expected increase in premiums from current premiums. This is because the group credit life insurance premiums rise as the ages of those borrowers in the insured group rise. Premium payments also rise with the ages of insured borrowers, in line with the seasoning of the loan pools. We assume increased premium payments in our cash flow analysis as set out in table 31.

In some cases, we may adjust the level of additional costs for group credit life insurance to better reflect a pool's characteristics and composition. For example, there might be cases where borrowers enter into new types of insurance that include wider coverage than typical insurance. In such cases, insurance premiums tend to be higher than existing ones, and therefore we might need to revise upward our assumptions for insurance premiums. In addition, there might be cases where some borrowers in the pool enter into life insurance while others do not. In such cases, we may assume premium payments for policyholders only.

Fixed-rate convertible loans

Many lenders in Japan often offer fixed-rate convertible mortgage loans to borrowers. In this type of mortgage loan, borrowers can choose either a fixed or floating rate when the initial fixed-rate period expires. We typically assume borrowers of such loans will opt to repeat the 10-year fixed-rate period in our cash flow analysis.

Details Of Further Considerations In The Surveillance Of Japanese RMBS Or Covered Bond Transactions

The characteristics of a mortgage loan pool evolve over time. For instance, some loans in a pool will default or prepay, and other characteristics of the pool will change. As a result, features initially consistent with the archetypal pool may deviate from those of the archetype over a pool's life.

Ongoing surveillance of the revised characteristics of individual mortgage loans allows the assessment of risks in less homogenous pools.

Rating changes resulting from surveillance emanate from changes in the performance of a pool and the analysis of pool performance data and trends. Changes in the ratings on supporting counterparties and in the rating on the sovereign can also influence our ratings on the notes.

For a given Japanese RMBS portfolio with changed pool characteristics, the surveillance analysis would use the same methodology and assumptions described in this document.

If loan-by-loan data are available only at the time of a transaction's closing, the transaction is exposed to a risk of the undetected changes in its pool composition during the surveillance period. For example, the weighted-average LTV of a mortgage pool containing loans with a wide distribution of original LTVs at closing may rise because of prepayments of loans with lower-than-average LTVs. This leads to possible adverse selection risk as well as an associated risk coming from a lack of information about which loans are outstanding in the pool and outstanding loan balance for transactions lacking updated loan-by-loan data in the surveillance period. We use the approaches below to address this risk.

• We apply a hypothetical current loan balance reflecting the scheduled amortization as the current loan balance at the time of the surveillance.
• We do not apply seasoning adjustments.
• We apply additional originator adjustments until the weighted loan seasoning of the pool exceeds five years.
• We capture a stress related to the pool level arrear in the originator adjustment.

For apartment loan RMBS, we always receive updated loan-by-loan data during surveillance, considering lower loan counts in the pool compared with transactions backed by residential or condominium investment loans.

CRITERIA SUPPLEMENT: KOREA

Variables Used As An Input In Our Calculation Of The Foreclosure Frequency

Korea Mortgage Market Assessment (MMA)

Our MMA in a given jurisdiction consists of an assessment on a six-point scale ranging from "very low risk" to "extremely high risk".

Table 36

Mortgage market assessment
Korea MMA	Low risk
Sub factors
Economic risk	Low risk
Industry risk	Intermediate risk
Lender - recourse assessment	Full recourse

As per our MMA Methodology Framework (see Appendix I and table 5 of our residential mortgage criteria). The MMA for Korea reflects our view of:

• The risks in the Korean banking system;
• Relative to the overall Korean banking system, lower credit risk and stronger and more stable profitability for mortgage loans, compared with corporate loans and other household loans;
• Low unemployment sensitivity to changes in economic output in Korea, combined with a relatively less supportive social welfare system, relative to some countries with more established systems; and
• Borrowers' willingness to pay in low/negative equity scenarios because of the full-recourse feature typically associated with standard Korean mortgage products provided by bank lenders.

The Korea archetypal foreclosure frequency anchors

When applying paragraphs 15, and 26, and table 1 of our residential mortgage criteria, based on an MMA of "low risk" ('2'), we set Korea's foreclosure frequency assumption as indicated in table 37 below. As per paragraphs 23-25 of our residential mortgage criteria, the 'B' foreclosure frequency assumption reflects our assessment of historical performance and our expectations for future performance over the medium term, given anticipated macroeconomic conditions. In our analysis, we considered default expectations across mortgage portfolios and the performance of outstanding Korean covered bond programs and transactions we rate, in both cases adjusted for seasoning to reflect expected lifetime defaults.

Table 37

Foreclosure frequency anchors
Rating level*	Archetypal foreclosure frequency (%)
AAA	10.0
AA	7.5
A	5.0
BBB	3.2
BB	2.1
B	1.1
*Assumptions for intermediate rating levels are interpolated.

Korea archetypal pool

We define the archetypal pool for Korea as follows:

Table 38

Korea archetypal pool
Characteristics by type	Archetypal features
Pool
Pool size	At least 250 loans at issuance
Originator	No adjustment factor related to the quality of the lender's underwriting or for features not specifically covered by other adjustments.
Geographic distribution	Diversified nationally
Borrower
Borrower type	Borrower is a private individual
Citizenship	Korean citizen
Employment type	Not self-employed or unemployed
Performance status	Not delinquent.
Borrower credit history	No adverse credit history
Affordability	Lender has assessed the borrower's income
Loan
Currency/denomination	KRW
Seasoning	Up to 60 months
Loan amortization profile	Fully amortizing
Loan product	Full recourse loans with no payment shock or interest only to maturity feature.
Term to maturity	No greater than 35 years
Interest rate	Fixed-rate, fixed-reset, or floating-rate loans
Security	First-lien mortgage on the property. Super-priority rights arising from key money deposits placed by small lease tenants rank prior to the first-lien mortgage claim under Korean regulations if conditions are met.
Loan purpose	Purchase or to refinance the balance on an existing loan (where the lender has fully re-underwritten the loan) of a residential property for owner occupation.
Loan-to-value (LTV)	70% (Calculated weighting the original LTV and current indexed LTV in an 80:20 ratio for full recourse loan, and 50:50 ratio for limited recourse loan).
Debt-to-income (DTI)	40% total debt DTI
Property
Property type	Residential
Occupancy status	Owner occupied and primary residence
Valuation method	Full valuations on the mortgaged property from a real estate appraiser, valuations calculated based on transaction-based database provided by regulator-recognized entities, or market-value provide by tax administrations.
Valuation amount	Up to the applicable jumbo valuation threshold specified in table 41.

Variables Considered When Calculating The Foreclosure Frequency Of A Mortgage Pool

For each characteristic and attribute that differs from the archetype at a pool, loan, or borrower level, the criteria apply a corresponding adjustment to the foreclosure frequency. The following table shows the applicable variables.

Table 39

Variables considered when calculating the foreclosure frequency of a mortgage pool
Originator adjustment	Typically 0.7x-1.3x or higher applied at pool level. For country-specific originator adjustment factors, see "Originator adjustment" in "Details Of Further Considerations When Calculating The Foreclosure Frequency Of A Korean Mortgage Pool" below. In particular, we may apply an originator adjustment to capture missing loan-level data for self-employed borrowers or loans that are backed by second homes, or investment properties where the loan is originated against the borrower's income. In sizing this adjustment, for pools that we consider to be representative of the overall mortgage market, we typically assume 25% of loans are self-employed borrowers, and 20% of loans are backed by second homes.
LTV	Type 2 LTV curve for full-recourse loans; and Type 1 LTV curve for limited-recourse loans (see chart 7 below).
Combined LTV definition	The LTV is calculated weighting: 80% of the OLTV and 20% of the CLTV for full recourse loans; and 50% of the OLTV and 50% of the CLTV for limited recourse loans.
Loan affordability: Debt-to-income (DTI)	0.8x if DTI <= 28%; 1.0x if DTI = 40%; and 1.5x if DTI >= 70%. This adjustment is a continuous function (see chart 8 below). For borrowers without proven-income, higher of DTI and income documentation adjustment factor is applied. Only non-investment loan will be subject to the DTI adjustment.
Income documentation	Typically 1.5x adjustment for self-certification and missing data, and 1.25x for borrowers with alternative income proof; both removed gradually when loan is performing or in arrears less than 30 days, and seasoning exceeds 12 months as follows: <=12: 100%; 12-24: 85%; 24-36: 80%; 36-48: 55%; 48-60: 35%; ` 60-72: 15%; and >72: 0. For borrowers without proven-income, higher of DTI and income documentation adjustment factor is applied. Only non-investment loan will be subject to the income documentation adjustment.
Seasoning (adjustment factors for loan seasoning)	1.0x for seasoning <=5 years; 0.75x for seasoning >5 and <=6 years; 0.70x for seasoning >6 and <=7 years; 0.65x for seasoning >7 and <=8 years; 0.60x for seasoning >8 and <=9 years; 0.55x for seasoning >9 and <=10 years; and 0.50x for seasoning >10 years. Factor applies only to loans that are current or in arrears less than 30 days.
Employment status	1.0x for regular salaried employee; 1.25x for self-employed person and irregular salaried employee or contractor; 1.5x for pensioner and retired borrowers; and 2.0x for unemployed person or other employment status. Only non-investment loan will be subject to the employment status adjustment.
Second and subsequent lien loans	1.3x: Where we consider that the loan does not have significant risk layering, or where the second and subsequent lien was not taken out to consolidate debt (is akin to a further advance), and where there is data relating to the senior lien holder; 1.5x: Where the borrower is using the second and subsequent lien for consumption or consolidation of debt and where we consider that there is risk layering; and 1.7x: Where there is insufficient data to back up other second and subsequent lien adjustments.
Borrower occupancy status: investment property (buy-to-let)	1.7x for full recourse loans underwritten to non-borrower income, such as rental income from the investment properties.
Borrower occupancy status: owner-occupied (second home)	1.3x for second home and loans identified with investment purpose underwritten to borrower income if any.
Loan purpose	1.1x for refinancing unless the lender has undertaken a full re-underwriting procedure; and 1.2x for debt consolidation, equity release/cash out loans and other loan purpose. The adjustment does not apply if the loan is also second or subsequent lien.
Loan product type: payment shock and interest only	1.0x for fixed rate, fixed reset and floating rate; 1.2x for fixed to floating rate, removed six months after the end of fixed rate period; 1.2x for teaser rate, removed six months after the end of teaser rate period; 1.2x for IO loans, removed six months after the end of IO period; 1.2x for incremental repayment; 1.3x for balloon loans; 1.5x for bullet loans (interest-only to maturity); and 2.0x for negative amortization.
Loan product type: limited recourse	1.5x for limited recourse loan products, in addition to the loan product adjustments above.
Arrears	2.5x for loans currently 30-59 days delinquent; 5.0x for loans currently 60-89 days delinquent; and 100% foreclosure frequency for loans currently 90 days or more delinquent.
Reperforming loans	We typically apply adjustments for reperforming loans when a portfolio contains a material portion of reperforming loans as follows: Months since last 90 days+ in arrears or restructure dates <=12: 2.0x; >12<=24: 1.5x; >24<=36: 1.2x; and >36: no adjustment. Reperforming loans definition: Typically, loan 90 or more days past due or restructured in the three years prior to the analysis date and is current as of that date.
Nonresidential use loans/commercial borrowers	1.5x-2.0x for commercial/mixed-used property; and 2.0x for commercial borrower. Limit of 40% of the pool at issuance.
Geographic concentration	1.20x adjustment applied to the excess above the regional concentration thresholds (% of pool balance in table 40 below).
Term to maturity	1.0x for loan term <=35 years; and 1.2x for loan term >35 years, unless performance data shows similar performance to loans with a shorter term
Residency status	Up to 2.5x adjustment factor for non-Korean citizens. We might apply a lower adjustment if the originator or servicer provides evidence that the origination process is robust and the performance of loans to non-Korean citizens is materially the same as that of loans to Korean citizens.

Table 40

Concentration limit for geographic concentration variable
Region	Concentration limit (%)
Seoul	40
Busan	14
Daegu	10
Incheon	12
Gwangju	6
Daejeon	6
Ulsan	5
Sejong	3
Gyeonggi-do	49
Gangwon-do	6
Chungcheongbuk-do	7
Chungcheongnam-do	9
Jeollabuk-do	8
Jeollanam-do	8
Gyeongsangbuk-do	11
Gyeongsangnam-do	14
Jeju	3

Details Of Further Considerations When Calculating The Foreclosure Frequency Of A Korean Mortgage Pool

Originator adjustment

Specific examples of the factors considered in determining the originator adjustment for a pool of Korean residential loans typically include but are not limited to:

• Origination and underwriting quality;
• Servicing quality;
• Originator specific performance to address idiosyncratic risks that are not captured in other adjustments;
• If we believe originator's LTV calculation is materially different from market standards or we believe inclusion of second and subsequent-ranking balances will significantly increase LTV level;
• If we believe the originator's practices on DTI calculation or requirements are significantly different from market standards, or the performance data suggests additional negative or positive adjustment;
• If we believe additional risk is not sufficiently captured in loan product adjustments;
• Unavailable data on loan and borrower characteristics (for example, LTV, DTI, income verification, restructurings, defaults, prior mortgage arrears, and property valuation method);
• If originators can't provide data at loan level for employment status, second homes or investment properties underwritten to borrower income, we apply an originator adjustment (see table 39). We typically size this originator adjustment by multiplying the applicable adjustment by our assumption of the proportion of loans with the relevant characteristic in the pool. We determine this assumption based on pool-level data, if available, the originator's underwriting policies, or market-level data, if representative;
• Loans to borrowers with negative credit histories;
• Recent changes in product offering, credit score process or underwriting guidelines, where the impact of which is not yet visible in performance metrics;
• Positive or negative selection not captured in other adjustments;
• Scenarios where the performance of a pool that has been sold deviates from our expectations for pools from that originator and for which a neutral originator adjustment is assumed;
• Weak representations and warranties of the loans in the transaction documentation and pool audit results;
• Dynamic or revolving asset pools;
• Any transaction's specific geographical concentration not captured by regional concentrations;
• We may analyze data from the issuer/servicer for reperforming loans on re-default rates stratified by forbearance type to calibrate the originator adjustment for such transactions; and
• Any other observed risk characteristics that are not adjusted for at a loan-by-loan level.
• For covered bonds, we may reduce the foreclosure frequency to reflect an issuer's willingness and ability to continue managing the cover pool. This assessment considers in particular the following factors:
• The existence in the transaction documentation of a periodic test of the pool's credit quality;
• The rating on the issuer;
• The importance of the program in the issuer's funding mix;
• The frequency of issuance from the program;
• The number of different covered bond programs that the issuer runs;
• Whether the issuer differentiates the way in which it manages the pools backing such covered bond; and
• The eligibility criteria for the cover pools.

The criteria set out that we consider the potential changes of credit risk over time. For RMBS transactions backed by a pool whose assets change (e.g., by virtue of loan substitutions, product switches, or similar, or revolving), in determining the pool's weighted-average foreclosure frequency (WAFF) and the weighted-average loss severity (WALS), we consider the potential increase of credit risk over time as a result of changes in pool composition. We assess possible deterioration in pool composition based on the transaction's documented asset-eligibility criteria, the history of the originator and, in particular, any observed changes in origination, underwriting, and related performance.

Loan to value (LTV)

LTV is calculated in a three-stage process. Stage 1: The original LTV (OLTV) is calculated using the original loan balance at the time of the latest advance, typically at loan origination and the property valuation at the time of that advance. Stage 2: The current LTV (CLTV) is calculated using the loan balance as of the portfolio cut-off date and the current indexed property value. The CLTV calculation incorporates any applicable valuation haircuts. Stage 3: The LTV is calculated by using the specific weighting of the OLTV and CLTV as described in table 39. We may also consider the maximum drawable balance, further advance, and purchase price if applicable in our analysis.

The OLTV and CLTV are based on the full loan balances secured on the property, including loan parts outside the asset pool and prior-ranking balances. CLTV could reflect updated prior-ranking balances if data is available. Second and subsequent-ranking balances will also be included in both OLTV and CLTV if data is available. Prior-ranking balances could include but are not limited to super-priority rights for small leases and first-lien "chonsei," if any. Chonsei are a type of security deposit by tenants common in the Korean real estate market. Second and subsequent-ranking balances could include but are not limited to second-lien chonsei.

In the LTV calculation, mortgage loan insurance covering super-priority rights for small leases in excess of the LTV limit set by the regulators is not considered. Please refer to 'Mortgage loan insurance' below.

An originator adjustment may apply if the originator's LTV calculation is materially different from market standards or we believe the inclusion of second and subsequent-ranking balances will significantly increase the portfolio LTV levels.

Loan affordability: Debt-to-income (DTI)

DTI is defined as total principal and interest payment for all household loans over income (following the debt-service ratio (DSR) definition by local regulators). Secondary income and spousal income might be aggregated if this is in compliance with the regulators' guidelines. If spousal income is considered, spousal debt will also be considered.

If loan-level data is not available for measuring affordability, we make adjustments based on our assessment of the originator's practices relative to market standards via the originator adjustment at the pool level.

We may apply originator adjustments in addition to the DTI adjustment on a case-by-case basis, depending on the performance data provided, or if we believe the originator's practices are significantly different from market standards.

Seasoning (adjustment factors for loan seasoning)

If corresponding data are available, the criteria may consider the seasoning of the performing relationship between the borrower and the mortgage lender of an existing mortgage loan being refinanced, instead of the seasoning of the new loan that results from such refinancing. For instance, a loan may have been refinanced with a new loan that has the same term or a shorter term, and the loan balance and interest rate are the same or lower. In that case, seasoning could be based on the origination date of the original loan rather than the date of the refinancing.

Second and subsequent lien loans

We apply an adjustment factor for second and subsequent lien mortgage loans as indicated in table 39. We differentiate a variable by the other lien loan's data availability.

Mortgage loans with super-priority rights for small leases are not considered as second lien as long as there are no other prior-ranking balances. We typically aggregate the super-priority rights for small leases with the mortgage claim and treat them as a single loan in the analysis.

We may consider different adjustments on a case-by-case basis, depending on the servicers' underwriting policies, track record, and performance data provided.

To the extent a pool exhibits a concentration of third and subsequent lien loans, then we may apply a higher originator adjustment.

In all cases above, the loan purpose adjustment does not apply.

Borrower occupancy status

We apply adjustment factors for loans to finance investment properties. The investment property classification includes second homes, investment property with loans underwritten to borrower income, and investment property with loans underwritten to non-borrower income, such as the property's DSCR.

Typically, for investment properties with loans underwritten to non-borrower income, borrower-related adjustments such as DTI, self-certification and employment status do not apply. We may consider different adjustments on a case-by-case basis, depending on the performance data provided.

Loan product type (payment shock and interest only)

Where a borrower may face payment shock from two separate sources, for example, a borrower with a fixed-to-float and incremental repayment loan, we do not apply the adjustment twice, but apply the highest adjustment once. However, for limited recourse loan products, we apply both the adjustment for limited recourse, and any further adjustment for payment shock and interest-only loans, where applicable.

We might apply additional adjustments based on product type through originator adjustments if we believe risks are not sufficiently captured in abovementioned adjustments.

Re-performing loans and restructured/modified loans

We may consider different adjustments on a case-by-case basis, depending on the servicers' restructure policies, track record, and performance data provided.

We typically define a reperforming loan as a loan that has been 90 or more days past due or restructured in the three years leading up to the analysis date and is current as of that date.

When a reperforming arrangement is made, it is typical that a full reassessment of the borrower's affordability capacity is made; we consider this akin to a re-underwriting of the loan. Accordingly, for pools classified as reperforming, we calculate potential future seasoning credit based on the date a loan was previously 90 or more days in arrears.

In addition to the original loan and borrower information provided, we may also consider updated data sourced through the restructuring process in our analysis of reperforming loans, where available, on a case-by-case basis.

In addition, as part of the analytical process, we analyze data from the issuer/servicer on re-default rates stratified by forbearance type. This analysis is used to calibrate the originator adjustment for such transactions.

Geographic concentration

If a pool has significant geographic concentration risk that we believe is not sufficiently captured, we may account for it using the originator adjustment.

Residency status

We would apply a lower adjustment if the originator or servicer provides evidence that the origination process is robust and the performance of loans to non-Korean citizens are comparable to Korean citizens.

Variables Considered When Calculating The Loss Severity Of A Mortgage Pool

The table below shows the variables we consider when calculating the loss severity of a mortgage pool.

Table 41

Variables considered when calculating the loss severity of a Korean mortgage pool
Property indexation	To estimate price fluctuations for each property since loan origination, we use the housing purchase price composite indices published by Kookmin Bank. We base property value assumptions on region-specific index data. We reflect changes in the index without adjustment (i.e., we reflect 100% of the changes in the index, irrespective of the direction).
Jumbo valuation threshold*	KRW900 million
Valuation haircuts	5% if the valuation is based on sales price or other valuation methods excluding below: 0% for full appraisal, valuation based on transaction-based database provided by regulator-recognized entities, or market-value provided by tax administrations.
Foreclosure costs§	KRW5 million fixed costs; and 3% variable as a percentage of the post-repo, jumbo-adjustment, commercial/mixed-use properties adjustment valuation.
Foreclosure timeline/period	18 months for property value below or equal to jumbo valuation threshold; and 24 months for property value exceeding the jumbo valuation threshold.
Commercial/mixed-use properties	1.15x adjustment to the MVD.
*We increase MVD assumptions for jumbo valuations. We apply an adjustment of 20% on the excess above the jumbo threshold. For example, for a property with an indexed-valuation of KRW1 billion we apply an adjustment on the difference between KRW1 billion and KRW900 million (jumbo valuation threshold). Jumbo valuation adjustment, the product of 20% x KRW100 million in this example, is then deducted from the indexed valuation after applying the post-repossession market value decline. §The post-repo, jumbo-adjustment, commercial/mixed-use properties adjustment valuation is the property value obtained after the application of the repossession market value decline at the relevant rating as well as the jumbo valuation adjustment and commercial/mixed-use properties adjustment to the indexed-valuation. **MVD--market-value decline. The MVD of a repossessed property (Repo MVD) is capped at 75%.

Details Of Further Considerations When Calculating The Loss Severity Of A Korean Mortgage Pool

Valuation haircut

A standard valuation method in Korea is through a transaction-based database provided by Kookmin Bank, Ministry of Land, Infrastructure and Transport, Korean Real Estate Board, or other government-recognized entities, because a full appraisal is not common in the Korean mortgage market for certain property types. Sometimes, the valuation can also be provided by the tax administration. Considering this, we generally do not apply a haircut to these valuation types. We apply a certain haircut if the valuation is based on sales price or other valuation methods (see table 41). Where relevant, we may make an adjustment to better reflect the pool's characteristics. For example, we may revise up or down a lender's valuation if they employ less or more conservative valuation methods than market standards.

A valuation haircut might apply at pool level when loan level data is not available or might vary when an originator's valuation method significantly deviates from the industry standard.

Forced-sale discount (FSD)

On a case-by-case basis, we may increase or decrease the forced-sale discount where there is sufficient information to support it being either higher or lower than envisaged using the standard calculation.

Accrued and unpaid interest

Prior-ranking balances, typically superpriority rights or first-lien chonsei, will not accrue interest and will not result in unpaid interest to be incorporated into a loan's loss severity if a cash flow analysis is not performed.

Mortgage loan insurance

To determine the loss severity, no credit is given to mortgage loan insurance which covers super-priority rights for small leases in excess of the LTV limit set by the regulators. We currently don't have any historical data for claims-payout ratio to determine an assumption in accordance with the methodology outlined in "Methodology For Assessing Mortgage Insurance And Similar Guarantees And Supports In Structured And Public Sector Finance And Covered Bonds," Dec. 7, 2014. We may consider giving credit to the mortgage loan insurance when more claims-payout ratio data is available to us.

Details Of Further Considerations When Calculating Minimum Credit Enhancement Levels In The Analysis Of A Korean RMBS Or Covered Bond Transaction

Table 42

Minimum credit enhancement
	AAA	AA	A	BBB	BB	B
CE floor*	4.00	2.50	1.50	1.00	0.50	0.35
Hard CE floor	2.50	2.00	1.50	N/A	N/A	N/A
*CE floor is applied to final CE results in cash flow analysis, instead of credit losses from credit model. The floor is calculated as a percentage of pool balance. N/A--Not applicable.

Variables considered in the cash flow model runs

Table 43

Variables considered in the cash flow model runs for a Korean mortgage pool
Default timing scenarios	Default timing curves: Typically, "front-loaded" and "back-loaded". The peak level of stress in each scenario is referred to as a recession for covered bonds. It corresponds to months 1-60 in both scenarios. (See table 44 below).
Interest rate risk	Standard run scenarios: Typically, "Up", "Up-down", "Down" and "Down-up" (see "Methodology To Derive Stressed Interest Rates In Structured Finance," Oct. 18, 2019). The curves vary by stress scenario. Our stressed interest rate scenarios for Korea are published in "Sector And Industry Variables: Methodology To Derive Stressed Interest Rates In Structured Finance," April 21, 2023.
Prepayment scenarios	Typically, "Low" and "High". See table 45 below.
Senior fees and expenses	Servicing fee: 40 bps per annum; and Extraordinary fee: up to 25 bps per annum.
Delinquency stress	Delinquency level: Two-third of the WAFF in each of the first 18 months; and Recovery timing: 18 months.

Table 44

Default timing curves (% of WAFF)
Number of years*	Front-loaded (%)	Back-loaded (%)
1	10	5
2	20	10
3	20	15
4	15	20
5	10	15
6	10	10
7	5	10
8	5	5
9	5	5
10		5
Total	100	100
*Percentage of weighted-average foreclosure frequency applied in each term. For modeling purposes, the percentage in each stage divided by 12 is applied each month. WAFF--Weighted average foreclosure frequency.

Table 45

Prepayment rates
	Low	High
CPR (%)	3	20
CPR--Constant prepayment rate.

Details Of Further Considerations When Applying The Cash Flow Criteria To The Analysis Of A Korean RMBS Or Covered Bond Transaction

In our rating analysis, we also assess a transaction's payment structure and cash flow mechanics. This analysis uses our own quantitative models to evaluate whether the cash flows from the assets suffice, at the applicable rating levels, for making timely payments of interest and ultimate payment of principal (i.e., by or before the legal maturity date).

In our cash flow analysis, we use the pool-level WAFF and WALS described in our residential mortgage criteria and previous sections as inputs, to reflect credit stress at each rating level.

During modeling, cash flow stresses test the credit and liquidity support the assets need to make timely interest payments and final principal payments by their legal maturity dates. Our models consider any available structural support, such as cash reserves, liquidity facilities, and hedging arrangements.

For revolving stand-alone RMBS structures (i.e., structures backed by a pool whose assets change or revolve), the modeling approach aims to reflect the structure after the activation of any "early amortization" (or stop-substitution) triggers and applies cash flow stresses from this point. An early amortization trigger is an event or situation that halts the substitution of assets in a revolving loan pool. We do not apply this approach to pools supporting covered bonds, even though the assets in these pools may be substituted over time. This is because the starting assumption of the collateral analysis under our covered bond criteria is the default of the issuing bank, as a result of which we do not expect the cover pool to be actively managed (it would become a static pool).

This criteria supplement does not include the assumptions used to assess refinancing costs in covered bonds structured with an asset-liability mismatch (such as target asset spreads). Those are described in "Covered Bonds Criteria," Dec. 9, 2014.

In our surveillance of existing ratings, cash flow modeling may show that under the 'B' stress, a particular tranche will miss interest payments or fail to repay the principal by or before the final legal maturity date. If this is the case, then our initial assessment, assuming all factors remain the same, may be to consider lowering the rating on those securities to 'B-' or lower.

Depending on our view of a transaction's immediate cash flow position, the rating could move into the 'CCC', 'CC', or 'C' rating category, consistent with our ratings definitions, and in accordance with "Criteria For Assigning 'CCC+', 'CCC', 'CCC-', And 'CC' Ratings," Oct. 1, 2012.

Default timing

The cumulative amount of defaults for cash flow modeling is equal to the pool's WAFF, whereas the cumulative amount of recoveries is calculated as (1 minus WALS).

In some cases, we may make an adjustment to the default timing curves to better reflect the pool's composition and transaction structure. For example, we may shorten the default curves where the mortgage portfolio's average remaining term gets shorter.

Foreclosure period assumptions represent the estimated time to repossess and sell a property upon a default (see table 41). They also reflect the typical time necessary for judicial proceedings and any other likely delay.

Cash flow modeling considers the negative carry resulting from interest due on the rated liabilities during the foreclosure period.

The loss severity estimates used in the cash flow modeling are based on the loan principal and assume no recovery of interest accrued on the mortgage loans during the foreclosure period.

Interest rate scenarios

Specific structural features may involve using additional cash flow stresses, such as alternative interest rate patterns or different default-timing curves, among others.

Prepayment scenarios

Residential loan prepayments vary the amount of excess spread available, and this may affect the absolute level of defaults exhibited in a transaction.

When analyzing the payment structure and cash flow mechanics of Korean RMBS and covered bonds, we typically test the transaction's ability to withstand high and low prepayment scenarios, as set out in table 45.

We may raise the prepayment assumptions if a pool's historical prepayment rates were higher than historical averages or if a transaction was particularly sensitive to prepayment risk. We may also reduce prepayment stress in situations where long-term historical data support lower prepayment rate assumptions for a specific loan product.

Modeling of reinvestment rates and spread compression

We typically do not assume any revenues associated with any cash or securities the issuer holds unless there is an eligible guaranteed investment contract (GIC) in place.

On a case-by-case basis, we apply a spread compression assumption if we identify a risk of material decrease in spread over time. For example, borrowers paying higher interest rates might tend to prepay earlier or indicate higher default risk.

Originator insolvency, commingling, and set-off

The analysis of any commingling or setoff risks that can result from an originator's or servicer's insolvency follows the application of the counterparty criteria (see "Counterparty Risk Framework: Methodology And Assumptions," March 8, 2019). The counterparty criteria determine the maximum supported rating based on available mitigation of such counterparty risks. Mitigants may include a counterparty's commitment to implement remedies upon a downgrade, or structural factors such as the coverage of the risk through credit enhancement, demonstrated through the modeling of any exposure.

Regarding commingling risk, the degree to which a collection account holder's insolvency affects the cash flow from the assets in a mortgage loan pool depends on the collection account's characteristics, if the collection account is not in the name of the RMBS issuer.

The amount at risk depends on the timing of scheduled payments from borrowers, the frequency of transfers into the transaction account, and the level of prepayments.

Modeling of senior fees and expenses

The modeling of an issuer's foreseeable expenses might use stressed costs to reflect the need to replace the initial service provider.

The most significant portion of senior fees and expenses is related to the servicer. Estimation of transaction servicing fees usually includes a certain increase from the amount stipulated in the transaction agreement to account for a potential replacement of the servicer. We also examine the degree to which the structure is typical, and the levels of servicing fees in the market. The estimate of servicing fees in table 43 reflects our findings.

To reflect the likely cost of replacing the initial servicer, we set servicing fees at the higher of the contractual rate and the assumed rates in table 43.

Other unexpected expenses may arise during the life of a rated security, such as when amending transaction documents, costs associated with perfection of title, and any other unexpected costs required for the orderly maintenance of the assets. In the absence of any structural mechanism, such as a reserve designated to cover substantial or full amount of the unexpected expenses, an expense cap, or if the cap is significantly lower than industry standard, we might assume extraordinary expenses of up to a certain amount in our cash flow analysis (see table 43).

We may apply different assumptions from those detailed in table 43. For example, we may lower the level to reflect easier/simpler tasks of the servicer. On the other hand, we may raise the level for an exotic transaction to reflect a difficulty of the servicer's replacement.

Delinquency

We assume a delay of a proportion of scheduled interest and principal receipts in each of the first several months of a hypothetical recession, and set full recovery of the arrears to take place several months after the delinquency occurs (see table 43). The cash flow stress for delinquencies is independent of the arrears adjustment to the WAFF.

For pools that contain residential loans with an option to temporarily suspend the periodic payments (payment holiday loans), the criteria supplementincludes a delay of a proportion of scheduled interest and principal receipts. In situations where there is the potential for payment holidays to be granted after a loan's inception or where payment holidays have been granted (e.g., due to governments' and banks' forbearance measures for households and small and midsize corporates), we may apply an additional stress in our cash flow analysis where relevant. In those instances, the criteria supplement includes a delay of a proportion of scheduled interest and principal receipts based on an estimate of the proportion of a pool that opts to take a payment holiday and the likely duration of the holiday. The likely duration will be assessed with reference to factors that may include, but are not limited to, relevant legislative frameworks, collateral credit quality, servicers' policies, and available servicer data on payment holidays granted.

CRITERIA SUPPLEMENT: NEW ZEALAND

Variables Used As An Input In the Credit Analysis Of Mortgage Pools

New Zealand mortgage market assessment (MMA)

The MMA for New Zealand is "low risk" ('2' on a scale of '1' to '6'). This is based on an economic risk score of '3', a mortgage industry risk score of '2', and our assessment of recourse available to lenders in the market. As per our MMA methodology framework (see Appendix I and table 5 of our residential mortgage criteria), the MMA for New Zealand reflects our view of:

• The risks in the New Zealand banking system;
• New Zealand's historical unemployment sensitivity to changes in economic output, combined with a relatively supportive social welfare system; and
• Borrowers' willingness to pay in low/negative equity scenarios due to full-recourse.

New Zealand archetypal pool

We define the archetypal pool for New Zealand as follows:

Table 46

New Zealand--archetypal pool
Characteristics by type	Archetypal features
Pool
Pool size	At least 250 loans at issuance.
Originator	No adjustment factor related to the quality of the lender's underwriting or for features not specifically covered by other adjustments.
Geographic distribution	Diversified nationally.
Borrower
Borrower type	Borrower is a private individual and not a first-time buyer.
Employment type	Pay-as-you-go (PAYG) full time and part time.
Performance status	Not delinquent.
Borrower credit history	No adverse credit history.
Affordability	Lender has assessed and fully verified the borrower's income.
Citizenship	New Zealand resident.
Loan
Currency	New Zealand dollar
Seasoning	Up to 60 months.
Loan amortization profile	Fully amortizing.
Loan product	No payment shock feature.
Loan purpose	Purchase or to refinance the balance on an existing loan (where the lender has fully re-underwritten the loan) of a residential property for owner occupation.
Loan-to-value (LTV)	75% (calculated weighting the original LTV and current indexed LTV in an 80/20 ratio).
Net surplus ratio (NSR) or Debt-to-income (DTI)	Lender has assessed the borrower's affordability commensurate with market standards.
Security	First-lien mortgage on the property, or crown leaseholds with a lease term of at least 15 years longer than the loan term.
Interest rate	Fixed-rate or floating-rate loans.
Term to maturity	>=30 and <31 years.
Property
Property type	Residential.
Occupancy status	Owner occupied.
Valuation method	Full valuation (or appraisal) of mortgaged property by registered valuers at the time of loan approval.
Valuation amount	Up to NZ$2 million.

New Zealand 'AAA' and 'B' foreclosure frequency anchors

Based on an MMA of "low risk" ('2'), we set New Zealand's 'AAA' foreclosure frequency anchor for the archetypal pool at 10%, and the current 'B' foreclosure frequency assumption at 1.1%.

The 1.1% 'B' foreclosure frequency assumption reflects our assessment of historical performance and our expectations for future performance over the medium term, given likely macroeconomic conditions. In our analysis, we consider default expectations across mortgage portfolios and the performance of outstanding New Zealand RMBS transactions we rate, in both cases adjusted for seasoning.

New Zealand archetypal foreclosure frequency anchors

The following table shows the archetypal foreclosure frequency anchors.

Table 47

Archetypal Foreclosure Frequency Anchors
Rating level*	Archetypal foreclosure frequency (%)
AAA	10.0
AA	7.5
A	5.0
BBB	3.2
BB	2.1
B	1.1
*Assumptions for intermediate rating levels are interpolated.

Variables Considered When Calculating the Weighted-Average Foreclosure Frequency (WAFF) Of A Mortgage Pool

New Zealand foreclosure frequency adjustment factors for variations from the archetypal pool

Table 48

Foreclosure Frequency Adjustments
Factor	Adjustment to foreclosure frequency
Loan to value (LTV)	Type 2 LTV curve (full-recourse market). The LTV is calculated weighting 80% of the original LTV (OLTV) and 20% of the current indexed LTV (CLTV).
Loan affordability	0.95x if lender assessment more conservative than industry peers, and 1.10x if lender assessment is below industry peers.
Seasoning (adjustment factors for loan seasoning)	0.75x for seasoning >5 and <=6 years; 0.70x for seasoning >6 and <=7 years; 0.65x for seasoning >7 and <=8 years; 0.60x for seasoning >8 and <=9 years; 0.55x for seasoning >9 and <=10 years; and 0.50x for seasoning >10 years.   The adjustment applies only to loans that are current or in arrears for up to 30 days. The adjustment does not apply to loans with bullet, balloon, interest-only, or negative amortization features during their period of non-amortization.
Employment status	1.5x for non-pay-as-you-go (PAYG) full time & part time (including PAYG-casual, commission-based, pension, over 65, and retired); 1.5x for self-employed of <=2 years; 1.2x for self-employed of >2 years to <=5 years; and 4.0x for unemployed.
Borrower occupancy status	1.1x for investment loans or second homes
Loan purpose	1.2x for refinance with debt consolidation, or refinance with cash out/equity withdrawal.
Payment shock (Loan product type)	1.2x for teaser rate, during and up to six months after end of teaser rate period; 3.0x for balloon loans (partial amortizing with a residual payment), or bullet loans (interest only until maturity), or negative amortization loans; and 1.1x-2.0x for partially amortizing loans (interest only (IO) for a period, then revert to fully amortizing) as specified in table 49.
Arrears	2.5x for loans currently 30-59 days delinquent; 5.0x for loans currently 60-89 days delinquent; and 100% foreclosure frequency for loans currently 90 days or more delinquent.
Residency status	1.5x for non-New Zealand residents
Credit history and arrears history for non-conforming borrowers	2.5x for borrowers with one adverse credit history event; and 3.0x for borrower with two or more adverse credit history events.   The number of adverse credit history events is counted in reference to the number of credit events, not the number of defaults and judgements, within the past five years of NZ$1,500 or higher (*).   1.1x for borrowers with two events of arrears in the past 12 months; 1.2x for borrowers with three events of arrears in the past 12 months; 1.5x for borrowers with four events of arrears in the past 12 months; and 2.0x for borrowers with five or more events of arrears in the past 12 months.   For borrowers with both adverse credit history and arrears history, only the credit history adjustment factor applies. For borrowers with no adverse credit history but who have arrears history, only the arrears history adjustment factor applies.
Loan term	0.7x for loan term <30 years; and 1.2x for loan term >=31 years.   This adjustment factor does not apply to loans with IO periods, bullet loans, negative amortizing loans, or balloon loans.
Geographic concentration	1.2x applied to the exposure in excess of the region concentration limits; 1.5x applied to the exposure in excess of nonmetropolitan concentration limits; and 1.5x applied to the exposure in excess of postcode concentration limits.   See "Geographic concentration" section below for concentration limits.
Originator adjustment	Typically 0.7x-1.3x or higher.
Income verification for self-employed and low documentation loans	1.5x for no credible sources; 1.4x for one credible source; 1.35x for two credible sources; 1.3x for three credible sources; 1.25x for four credible sources; and 1.0x if tax returns are included. Adjustment factor reduces over a six-year period as specified in table 52.
First-time buyer	1.1x to first-time borrowers with no credit history, until a payment history of at least 18 months has been established, upon which the adjustment factor is no longer applied if the loan is less than 30 days in arrears.
Redraws and further advances	Typically 1.05x for redraw only or for further advance only, and 1.1x for both redraw and further advance.   The size of these adjustment factors is based on the capacity for redraw and further advances in a pool or transaction based on loan characteristics and documented restrictions on allowable further advances. Higher adjustment may apply if the pool or transaction structure has greater capacity for redraws or further advances.
(*) We consider a credit event to be one event in a borrower's life that may drive a number of events (such as one loss of employment or ill health) that leads to a borrower not being able to meet payment obligations.

Loan to value (LTV)

LTV is calculated in a three-stage process:

• Stage 1: The original LTV (OLTV) is calculated using the original loan balance and the property valuation at the time of loan origination.
• Stage 2: The current LTV (CLTV) is calculated using the loan balance as of the portfolio cut-off date and the current indexed property value. The CLTV calculation incorporates any applicable valuation haircuts.
• Stage 3: The LTV is calculated by taking 80% of the OLTV and 20% of the CLTV. We may also consider the maximum drawable balance or further advance if applicable in our analysis.

The OLTV and CLTV are based on the full loan balances secured on the property, including loan parts outside the asset pool and prior- and second-ranking balances, if any. The use of an 80/20 combination of OLTV and CLTV maintains comparability to other similar jurisdictions under our global framework.

A loan's CLTV is the current balance of a loan, reflecting the actual principal pay down, divided by the indexed initial value.

The continuous function shown in the equation below defines the adjustment factors in chart 9.

These figures are rounded to four decimal places for ease of readability.

Loan affordability

New Zealand lenders generally assess affordability by setting a stressed interest rate floor and adopt a dollar net surplus, uncommitted income, or equivalent measure. Although approaches are relatively uniform, the composition and derivation of income and expenses, stressed interest rate floor, and the maximum or minimum limits (whichever is applicable) could vary across originators. Hence any meaningful deviation from market standards by originators in their assessment of affordability could attract a positive or negative foreclosure frequency adjustment.

Payment shock (loan product type)

Although interest-only loans that revert to fully amortizing loans after the interest-only period (IO-term) expires somewhat mitigate refinancing risk, the interest-only feature can create a payment shock when the payments revert to fully amortizing over the remaining term of the loan (PI-term). The amount of the adjustment depends on the degree of payment shock in each case and is a product of the IO-term related adjustment factor and PI-term related adjustment factor in table 49. This reflects our view that the longer the IO-term relative to the PI-term, the higher the risk of payment shock-related default.

Table 49

Adjustments For Interest-Only (IO), Then Reverting To Fully Amortizing (PI) Loans
IO term (years)	<=5	>5 to <=10	>10 to <=15	>15 to <=20	>20 to <=25
Adjustment factor	1.1x	1.25x	1.5x	1.75x	2.0x
PI term (years)	<3	=3 to <5	=5 to <10	=10 to <15	=15 to 30
Adjustment factor	1.75x	1.5x	1.25x	1.1x	1.0x

Originator adjustment

The calculation of foreclosure frequency includes an originator adjustment. Specific examples of the factors considered in determining the originator adjustment for a pool of New Zealand residential loans typically include but are not limited to:

• The quality of origination and underwriting. Our review assesses a range of issues including the business and lending strategy; the governance and management structure; risk-management, compliance, quality assurance (including data quality and internal control framework); policies, procedures and training programs; the role of mortgage originators and brokers in the origination and underwriting process; the underwriting standards and the quality of credit review (including whether credit checks have been obtained, verification of savings history, quality of valuation policy); the frequency of exceptions to established underwriting guidelines; general insurance coverage over security properties; and the role of any mortgage insurer in the underwriting process (if applicable).
• The servicing quality. This includes a review of the loan servicing philosophy and experience, the experience in servicing assets of the type to be securitized, the quality of the loan-servicing platform, the financial strength of the servicer, the experience of management and staff, the role of mortgage originators and brokers, the risk management, quality assurance, compliance and internal control framework, arrears management and recoveries, the delinquency historical performance, the transaction processing and cash management, and investor reporting;
• The historical performance of an originator's loans relative to other originators.
• Where insufficient data are available on the borrower's employment status, S&P Global Ratings typically assumes a percentage of the pool is to self-employed borrowers, generally 25% of all full-documentation loans, and all low- and no-documentation loans. We may assume a different level to this for full-documentation loans where line-by-line data is not available, if we are of the opinion that the characteristics of the originator or the borrowers would suggest a different percentage is more appropriate. We apply a stress of 100% for low- and no-documentation loans because such loans are almost always provided to self-employed borrowers.
• Where the borrower's loan purpose is for refinancing and there is insufficient data available on whether the refinancing is with or without debt consolidation or cash out/equity withdrawal, we may assume a portion of the pool is used for refinancing with debt consolidation or cash out/equity withdrawal. Such an assessment considers the lender's underwriting policies and practices, the lender's portfolio parameters, and industry trends and averages;
• Where insufficient data is available on the borrower's first-time home buyer status, we typically assume a percentage of the pool is made to first-time home buyers, generally within a range of 10%-30%. We may assume a higher level than this where line-by-line data is not available if we are of the opinion that the characteristics of the originator or the borrowers would suggest a higher percentage is more appropriate.

Geographic concentration

We apply adjustment factors to the exposure in excess of the specific thresholds for each state, nonmetropolitan, and postcode shown in the table below.

Table 50

Geographic Concentration Limits By Region, Nonmetropolitan And Postcode
Region	Limit (%)
Auckland	<=60
Wellington	<=50
Canterbury	<=25
Otago	<=10
Nonmetropolitan	Limit (%)
Nonmetropolitan	<=10
Postcode	Limit (%)
Postcode	<=5

Classifications of postcodes in New Zealand are shown in the table below.

Table 51

New Zealand metro post codes*
PostCode	Region	Location
600	Auckland	Metro
602	Auckland	Metro
604	Auckland	Metro
610	Auckland	Metro
612	Auckland	Metro
618	Auckland	Metro
620	Auckland	Metro
622	Auckland	Metro
624	Auckland	Metro
626	Auckland	Metro
627	Auckland	Metro
629	Auckland	Metro
630	Auckland	Metro
632	Auckland	Metro
930	Auckland	Metro
931	Auckland	Metro
932	Auckland	Metro
1010	Auckland	Metro
1011	Auckland	Metro
1021	Auckland	Metro
1022	Auckland	Metro
1023	Auckland	Metro
1024	Auckland	Metro
1025	Auckland	Metro
1026	Auckland	Metro
1041	Auckland	Metro
1042	Auckland	Metro
1050	Auckland	Metro
1051	Auckland	Metro
1052	Auckland	Metro
1060	Auckland	Metro
1061	Auckland	Metro
1062	Auckland	Metro
1071	Auckland	Metro
1072	Auckland	Metro
2010	Auckland	Metro
2012	Auckland	Metro
2013	Auckland	Metro
2014	Auckland	Metro
2022	Auckland	Metro
2023	Auckland	Metro
2024	Auckland	Metro
2025	Auckland	Metro
2102	Auckland	Metro
2103	Auckland	Metro
2104	Auckland	Metro
2105	Auckland	Metro
2110	Auckland	Metro
2112	Auckland	Metro
2113	Auckland	Metro
3110	Auckland	Metro
3112	Auckland	Metro
3116	Auckland	Metro
3200	Auckland	Metro
3204	Auckland	Metro
3206	Auckland	Metro
3210	Auckland	Metro
3214	Auckland	Metro
3216	Auckland	Metro
3218	Auckland	Metro
5010	Wellington	Metro
5011	Wellington	Metro
5012	Wellington	Metro
5013	Wellington	Metro
5014	Wellington	Metro
5018	Wellington	Metro
5019	Wellington	Metro
5024	Wellington	Metro
5028	Wellington	Metro
6011	Wellington	Metro
6012	Wellington	Metro
6021	Wellington	Metro
6022	Wellington	Metro
6023	Wellington	Metro
6035	Wellington	Metro
6037	Wellington	Metro
8011	Canterbury	Metro
8013	Canterbury	Metro
8014	Canterbury	Metro
8022	Canterbury	Metro
8023	Canterbury	Metro
8024	Canterbury	Metro
8041	Canterbury	Metro
8042	Canterbury	Metro
8052	Canterbury	Metro
8053	Canterbury	Metro
8061	Canterbury	Metro
8062	Canterbury	Metro
9011	Otago	Metro
9012	Otago	Metro
9016	Otago	Metro
*All other postcodes are classified as non metro.

Income verification

A borrower's demonstrated loan repayment outweighs the level of income and asset verification at the time of loan origination, and the initial level of income and asset verification becomes less indicative of likelihood of default over time. To account for this, these criteria reduce the adjustment factors over a six-year period using the equation below, which incorporates a seasoning adjustment from 100% to 0% (see table 52) over the same period. Effectively, the documentation-related adjustment factor becomes neutral after six years.

1 + (initial documentation type adjustment – 1) x seasoning adjustment

Table 52

Percentage Of Documentation Adjustment Factors Applied By Loan Seasoning
Loan Seasoning	<=12	(12-24]	(24-36]	(36-48]	(48-60]	(60-72]	>72
Seasoning adjustment	100%	85%	80%	55%	35%	15%	0%

Variables Considered When Calculating The Weighted-Average Loss Severity (WALS) Of A Mortgage Pool

Table 53

Loss Severity (WALS)
Factor	Adjustment to loss severity calculation
Valuation haircut	5% if not a full valuation.
Property indexation	Based on residential property price data published by organizations or institutions we deem reputable.   We reflect 50% of index appreciation and 100% of index depreciation.
Over/undervaluation	See the "Assessment of property of over-/under-valuation" section of our residential mortgage criteria.
Property value	1.2x for property values of more than NZ$2 million and less than or equal to NZ$3 million; 1.225x for property values of more than NZ$3 million and less than or equal to NZ$4 million; 1.25x for property values of more than NZ$4 million and less than or equal to NZ$5 million; 1.275x for property values of more than NZ$5 million and less than or equal to NZ$6 million; and 1.3x for property values of more than NZ$6 million.
Foreclosure costs	NZ$10,000 fixed cost. 5% variable as a percentage of index-adjusted, post-repo, post valuation haircut if any, and post jumbo-adjusted valuation.
Foreclosure timeline/period	For property values of less than or equal to NZ$2 million, 12 months if located in metropolitan area, and 18 months if located in non-metropolitan area.   For property values of greater than NZ$2 million, 18 months if located in metropolitan area, and 24 months if located in non-metropolitan area.
Accrued and unpaid interest	No adjustment where cash flow analysis is performed; otherwise included in the loss severity calculation based on current interest rate through the foreclosure period.
The MVD of a repossessed property (Repo MVD) is capped at 75%.

Accrued and unpaid interest

Where the rating analysis includes cash flow modelling, the impact on a transaction's credit enhancement from accrued and unpaid interest on defaulted loans during the foreclosure period is incorporated into our cash flow analysis and is therefore not included in the loss severity calculation. Where the rating analysis does not include cash flow modelling, the accrued and unpaid interest is included in the loss severity calculation. In both situations, stressed interest rate assumptions is used.

Further Considerations

Lenders mortgage insurance

Some residential loans in New Zealand are insured for loss by lenders mortgage insurance. In such cases, we apply the methodology outlined in "Methodology For Assessing Mortgage Insurance And Similar Guarantees And Supports In Structured And Public Sector Finance And Covered Bonds," Dec. 7, 2014.

Borrower concentration analysis

In pools where concentrations to borrowers are evident, we may apply an additional minimum loss projection when determining the expected loss for a pool at the relevant rating level.

Pools may be concentrated to borrowers with large loan sizes relative to total pool size at close, or become concentrated as pools amortize. A loss on one or a small number of loans in a concentrated pool may have a disproportionate impact on credit enhancement, because the loss on specific large loans might be higher than the absolute level of credit enhancement.

To address the potential that greater losses could result if the loans with higher balances were to default, the criteria use the expected loss on the largest loan balances.

We typically apply the additional minimum loss projection when the top-10 borrower exposures exceed 10% of the total current pool balance.

The additional minimum loss projection estimations at each rating category are based on a certain number of loans defaulting and liquidating as outlined in table 54. To address the potential that greater losses could result if the loans with higher balances were to default, the criteria use the largest liquidation amounts for each rating category.

The minimum loss projection estimations at each rating level are based on a certain number of loans defaulting and liquidating as outlined in table 54, and are the higher of that derived using the RMBS credit criteria outlined above and that calculated as an additional minimum loss projection in accordance with this subsection.

Table 54

Rating Category Minimum Loss Projections For Concentrated Pools
Maximum potential rating	Minimum loss parameters
AAA	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'AAA' loss severity and the weighted-average loss severity at 'AAA' for the pool, plus the next nine largest loss exposures at the 'AAA' loss severity
AA category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'AA' loss severity and the weighted-average loss severity at 'AA' for the pool, plus the next seven largest loss exposures at the 'AA' loss severity
A category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'A' loss severity and the weighted-average loss severity at 'A' for the pool, plus the next five largest loss exposures at the 'A' loss severity
BBB category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'BBB' loss severity and the weighted-average loss severity at 'BBB' for the pool, plus the next three largest loss exposures at the 'BBB' loss severity
BB category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'BB' loss severity and the weighted-average loss severity at 'BB' for the pool, plus the next largest loss exposure at the 'BB' loss severity
B category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'B' loss severity and the-weighted-average loss severity at 'B' for the pool

We may also adapt the above approach to assess the features (such as arrears status) of particular loans that account for the top X loans (X being the number of loans as determined in accordance with table 54), to capture any potential for loss on those loans to be greater than that determined under the calculation outlined in table 54.

CRITERIA SUPPLEMENT: SINGAPORE

Variables Used As An Input In the Credit Analysis Of Mortgage Pools

Singapore mortgage market assessment (MMA)

The Mortgage Market Assessment (MMA) for Singapore is "low risk" ('2' on a scale of '1' to '6'). This is based on an economic risk score of '2', a mortgage industry risk score of '1', and our assessment of recourse available to lenders in the market. As per our MMA methodology framework (see Appendix I and table 5 of our residential mortgage criteria), the MMA for Singapore reflects our view of:

• The risks in the Singapore banking system;
• Singapore's low unemployment sensitivity to changes in economic output, combined with a relatively less supportive social welfare system relative to some countries with more established systems, a significantly higher level of savings and access to retirement savings system to meet mortgage repayments; and
• Borrowers' willingness to pay in low/negative equity scenarios due to full-recourse.

Singapore archetypal pool

We define the archetypal pool for Singapore as follows:

Table 55

Singapore--Archetypal Pool
Characteristics by type	Archetypal features
Pool
Pool size	At least 250 loans at issuance.
Originator	No adjustment factor related to the quality of the lender's underwriting or for features not specifically covered by other adjustments.
Geographic distribution	Diversified nationally.
Borrower
Borrower type	Borrower is a private individual and not a first-time buyer.
Employment type	Salaried full time and part time.
Performance status	Not delinquent.
Borrower credit history	No adverse credit history.
Affordability	Lender has assessed and fully verified the borrower's income.
Citizenship	Singapore resident.
Loan
Currency	Singapore dollar
Seasoning	Up to 60 months.
Loan amortization profile	Fully amortizing.
Loan product	No payment shock feature.
Loan purpose	Purchase or to refinance the balance on an existing loan (where the lender has fully re-underwritten the loan) of a residential property for owner occupation.
Loan-to-value (LTV)	75% (calculated weighting the original LTV and current indexed LTV in an 80/20 ratio).
Mortgage servicing ratio (MSR) or total debt servicing ratio (TDSR)	Reflect regulatory standard.
Security	First-lien mortgage on the property.
Interest rate	Fixed-rate or floating-rate loans.
Term to maturity	>=30 and <31 years.
Property
Property type	Residential.
Occupancy status	Owner occupied.
Valuation method	Full valuation (or appraisal) of mortgaged property by registered valuers at the time of loan approval.
Valuation amount	Up to S$1.8 million.

Singapore 'AAA' and 'B' foreclosure frequency anchors

Based on an MMA of "low risk" ('2'), we set Singapore's 'AAA' foreclosure frequency anchor for the archetypal pool at 10%, and the current 'B' foreclosure frequency assumption at 1.1%.

The 1.1% 'B' foreclosure frequency assumption reflects our assessment of historical performance and our expectations for future performance over the medium term, given likely macroeconomic conditions. In our analysis, we consider default expectations across mortgage portfolios and the performance of outstanding Singapore covered bond programs we rate, in both cases adjusted for seasoning.

Singapore archetypal foreclosure frequency anchors

The following table shows the archetypal foreclosure frequency anchors.

Table 56

Archetypal Foreclosure Frequency Anchors
Rating level*	Archetypal foreclosure frequency (%)
AAA	10.0
AA	7.5
A	5.0
BBB	3.2
BB	2.1
B	1.1
*Assumptions for intermediate rating levels are interpolated.

Variables Considered When Calculating the Weighted-Average Foreclosure Frequency (WAFF) Of A Mortgage Pool

Singapore foreclosure frequency adjustment factors for variations from the archetypal pool

Table 57

Foreclosure Frequency Adjustments
Factor	Adjustment to foreclosure frequency
Loan to value (LTV)	Type 2 LTV curve (full-recourse market). The LTV is calculated weighting 80% of the original LTV (OLTV) and 20% of the current indexed LTV (CLTV).
Loan affordability	No adjustment.
Seasoning (adjustment factors for loan seasoning)	0.75x for seasoning >5 and <=6 years; 0.70x for seasoning >6 and <=7 years; 0.65x for seasoning >7 and <=8 years; 0.60x for seasoning >8 and <=9 years; 0.55x for seasoning >9 and <=10 years; and 0.50x for seasoning >10 years.   The adjustment applies only to loans that are current or in arrears for up to 30 days. The adjustment does not apply to loans with bullet, balloon, interest-only, or negative amortization features during their period of non-amortization.
Employment status	1.5x for non-salaried full time & part time (including salaried-casual, commission-based, pension, over 65, and retired); 1.5x for self-employed of <=2 years; 1.2x for self-employed of >2 years to <=5 years; and 4.0x for unemployed.
Borrower occupancy status	1.1x for investment loans or second homes
Loan purpose	1.2x for refinance with debt consolidation, or refinance with cash out/equity withdrawal.
Payment shock (Loan product type)	1.2x for teaser rate, during and up to six months after end of teaser rate period; 3.0x for balloon loans (partial amortizing with a residual payment), or bullet loans (interest only until maturity), or negative amortization loans; and 1.1x-2.0x for partially amortizing loans (interest only (IO) for a period, then revert to fully amortizing) as specified in table 58.
Arrears	2.5x for loans currently 30-59 days delinquent; 5.0x for loans currently 60-89 days delinquent; and 100% foreclosure frequency for loans currently 90 days or more delinquent.
Residency status	1.5x for non-Singaporean residents
Credit history and arrears history for non-conforming borrowers	2.5x for borrowers with one adverse credit history event; and 3.0x for borrowers with two or more adverse credit history events. The number of adverse credit history events is counted in reference to the number of credit events, not the number of defaults and judgements, within the past five years of S$1,500 or higher (*). 1.1x for borrowers with two events of arrears in the past 12 months; 1.2x for borrowers with two events of arrears in the past 12 months; 1.5x for borrowers with four events of arrears in the past 12 months; and 2.0x for borrowers with five or more events of arrears in the past 12 months. For borrowers with both adverse credit history and arrears history, only the credit history adjustment factor applies. For borrowers with no adverse credit history but who have arrears history, only the arrears history adjustment factor applies.
Loan term	0.7x for loan term <30 years; and 1.2x for loan term >=31 years. This adjustment factor does not apply to loans with IO periods, bullet loans, negative amortizing loans, or balloon loans.
Geographic concentration	1.2x applied to the exposure in excess of the region concentration limits; and 1.5x applied to the exposure in excess of postcode concentration limits. See "Geographic concentration" section below for concentration limits.
Originator adjustment	Typically 0.7x-1.3x or higher.
Income verification for self-employed and low documentation loans	1.5x for no credible source, 1.4x for one credible source, 1.35x for two credible sources, 1.3x for three credible sources, 1.25x for four credible sources, and 1.0x if tax returns included. Adjustment factor reduces over a six-year period as specified in table 60.
First-time buyer	1.1x to first-time borrowers with no credit history, until a payment history of at least 18 months has been established, upon which the adjustment factor is no longer applied if the loan is less than 30 days in arrears.
Redraws and further advances	Typically 1.05x for redraw only or for further advance only, 1.1x for both redraw and further advance. The size of these adjustment factors is based on the capacity for redraw and further advances in a pool or transaction based on loan characteristics and documented restrictions on allowable further advances. Higher adjustment may apply if the pool or transaction structure has greater capacity for redraws or further advances.
*We consider a credit event to be one event in a borrower's life that may drive a number of events (such as one loss of employment or ill health) that leads to a borrower not being able to meet payment obligations.

Loan to value (LTV)

LTV is calculated in a three-stage process:

• Stage 1: The original LTV (OLTV) is calculated using the original loan balance and the property valuation at the time of loan origination.
• Stage 2: The current LTV (CLTV) is calculated using the loan balance as of the portfolio cut-off date and the current indexed property value. The CLTV calculation incorporates any applicable valuation haircuts.
• Stage 3: The LTV is calculated by taking 80% of the OLTV and 20% of the CLTV. We may also consider the maximum drawable balance or further advance if applicable in our analysis.

The OLTV and CLTV are based on the full loan balances secured on the property, including loan parts outside the asset pool and prior- and second-ranking balances, if any. The use of an 80/20 combination of OLTV and CLTV maintains comparability to other similar jurisdictions under our global framework.

A loan's CLTV is the current balance of a loan, reflecting the actual principal pay down, divided by the indexed initial value.

The continuous function shown in the equation below defines the adjustment factors in chart 10.

These figures are rounded to four decimal places for ease of readability.

Loan affordability

In underwriting housing loans, lenders in Singapore are required to adhere to regulator-determined and mandatory affordability measurements with the use of mortgage servicing ratio (MSR) for purchase of Housing Development Board (HDB) properties or total debt servicing ration (TDSR) for purchase of private properties. The definitions of both MSR and TDSR are highly prescriptive and set a consistent and well-defined benchmark across lenders. As a result, no negative affordability adjustment factor is applied for Singapore.

Payment shock (loan product type)

Although interest-only loans that revert to fully amortizing loans after the interest-only period (IO-term) expires somewhat mitigate refinancing risk, the interest-only feature can create a payment shock when the payments revert to fully amortizing over the remaining term of the loan (PI-term). The amount of the adjustment depends on the degree of payment shock in each case and is a product of the IO-term related adjustment factor and PI-term related adjustment factor in table 58. This reflects our view that the longer the IO-term relative to the PI-term, the higher the risk of payment shock-related default.

Table 58

Adjustments For Interest-Only (IO), Then Reverting To Fully Amortizing (PI) Loans
IO term (years)	<=5	>5 to <=10	>10 to <=15	>15 to <=20	>20 to <=25
Adjustment factor	1.1x	1.25x	1.5x	1.75x	2.0x
PI term (years)	<3	=3 to <5	=5 to <10	=10 to <15	=15 to 30
Adjustment factor	1.75x	1.5x	1.25x	1.1x	1.0x

Originator adjustment

The calculation of foreclosure frequency includes an originator adjustment. Specific examples of the factors considered in determining the originator adjustment for a pool of Singapore residential loans typically include, but are not limited to:

• The quality of origination and underwriting. Our review assesses a range of issues including the business and lending strategy; the governance and management structure; risk-management, compliance, quality assurance (including data quality and internal control framework); policies, procedures and training programs; the role of mortgage originators and brokers in the origination and underwriting process; the underwriting standards and quality of credit review (including whether credit checks have been obtained, verification of savings history, quality of valuation policy); the frequency of exceptions to established underwriting guidelines; general insurance coverage over security properties; and the role of any mortgage insurer in the underwriting process (if applicable).
• The servicing quality. This includes a review of the loan servicing philosophy and experience, the experience in servicing assets of the type to be securitized, the quality of the loan-servicing platform, the financial strength of the servicer, the experience of management and staff, the role of mortgage originators and brokers, the risk management, quality assurance, compliance and internal control framework, arrears management and recoveries, the delinquency historical performance, the transaction processing and cash management, and investor reporting.
• The historical performance of an originator's loans relative to other originators.
• Where insufficient data are available on the borrower's employment status, S&P Global Ratings typically assumes a percentage of the pool is to self-employed borrowers, generally 25% of all full-documentation loans, and all low- and no-documentation loans. We may assume a different level to this for full-documentation loans where line-by-line data is not available, if we are of the opinion that the characteristics of the originator or the borrowers would suggest a different percentage is more appropriate. We apply a stress of 100% for low- and no-documentation loans because these loans are almost always provided to self-employed borrowers.
• Where the borrower's loan purpose is for refinancing and there is insufficient data available on whether the refinancing is with or without debt consolidation or cash out/equity withdrawal, we may assume a portion of the pool is used for refinancing with debt consolidation or cash out/equity withdrawal. Such an assessment considers the lender's underwriting policies and practices, the lender's portfolio parameters, and industry trends and averages.
• Where insufficient data is available on the borrower's first-time home buyer status, we typically assume a percentage of the pool is made to first-time home buyers, generally within a range of 10%-30%. We may assume a higher level than this where line-by-line data is not available if we are of the opinion that the characteristics of the originator or the borrowers would suggest a higher percentage is more appropriate.

Geographic concentration

We apply adjustment factors to the exposure in excess of the specific thresholds for each region and postcode shown in the table below.

Table 59

Geographic Concentration Limits By Region And Postcode
Region	Limit (%)
Central	<=50
West	<=50
Northeast	<=40
East	<=40
North	<=30
Postcode	Limit (%)
Postcode	<=5

Income verification

A borrower's demonstrated loan repayment outweighs the level of income and asset verification at the time of loan origination, and the initial level of income and asset verification becomes less indicative of the likelihood of default over time. To account for this, these criteria reduce the adjustment factors over a six-year period using the equation below, which incorporates a seasoning adjustment from 100% to 0% (see table 60) over the same period. Effectively, the documentation-related adjustment factor becomes neutral after six years.

1 + (initial documentation type adjustment – 1) x seasoning adjustment

Table 60

Percentage Of Documentation Adjustment Factors Applied By Loan Seasoning
Loan Seasoning (months)	<=12	(12-24]	(24-36]	(36-48]	(48-60]	(60-72]	>72
Seasoning adjustment (%)	100	85	80	55	35	15	0

Variables Considered When Calculating The Weighted-Average Loss Severity (WALS) Of A Mortgage Pool

Table 61

Loss Severity (WALS)
Factor	Adjustment to loss severity calculation
Valuation haircut	5% if not a full valuation.
Property indexation	Based on Housing Development Board (HDB) resale price index as published by the Singapore government.   We reflect 50% of index appreciation and 100% of index depreciation.
Over/undervaluation	See the "Assessment of property of over-/under-valuation" section of our residential mortgage criteria.
Property value	1.2x for property values of more than S$1.8 million and less than or equal to S$2.3 million; 1.225x for property values of more than S$2.3 million and less than or equal to S$2.8 million; 1.25x for property values of more than S$2.8 million and less than or equal to S$3.3 million; 1.275x for property values of more than S$3.3 million and less than or equal to S$3.8 million; and 1.3x for property values of more than S$3.8 million.
Property type	1.2x for non-HDB properties.
Foreclosure costs	S$5,000 fixed cost. 5% variable as a percentage of index-adjusted, post-repo, post valuation haircut if any, and post jumbo-adjusted valuation.
Foreclosure timeline/period	18 months for property values of less than or equal to S$1.8 million, and 24 months for property values of greater than S$1.8 million.
Accrued and unpaid interest	No adjustment where cash flow analysis is performed; otherwise included in the loss severity calculation based on current interest rate through the foreclosure period.
The MVD of a repossessed property (Repo MVD) is capped at 75%.

Property type

HDB properties are foundational to the property sector in Singapore. Such properties are the dominant housing stock in Singapore with regulations and policies aimed at providing housing to Singapore citizens. Because we view HDB properties to be archetypical in our rating analysis, we therefore apply a 1.2x adjustment factor to properties that are non-HDB. This reflects the risk of greater price volatility of such non-HDB properties.

Accrued and unpaid interest

Where the rating analysis includes cash flow modelling, the impact on a transaction's credit enhancement from accrued and unpaid interest on defaulted loans during foreclosure period is incorporated into our cash flow analysis and is therefore not included in the loss severity calculation. Where the rating analysis does not include cash flow modelling, the accrued and unpaid interest is included in the loss severity calculation. In both situations, stressed interest rate assumptions is used.

Further Considerations

Borrower concentration analysis

In pools where concentrations to borrowers are evident, we may apply an additional minimum loss projection when determining the expected loss for a pool at the relevant rating level.

Pools may be concentrated to borrowers with large loan sizes relative to total pool size at close, or become concentrated as pools amortize. A loss on one or a small number of loans in a concentrated pool may have a disproportionate impact on credit enhancement, because the loss on specific large loans might be higher than the absolute level of credit enhancement.

To address the potential that greater losses could result if the loans with higher balances were to default, the criteria use the expected loss on the largest loan balances.

We typically apply the additional minimum loss projection when the top 10 borrower exposures exceed 10% of the total current pool balance.

The additional minimum loss projection estimations at each rating category are based on a certain number of loans defaulting and liquidating as outlined in table 62. To address the potential that greater losses could result if the loans with higher balances were to default, the criteria use the largest liquidation amounts for each rating category.

The minimum loss projection estimations at each rating level are based on a certain number of loans defaulting and liquidating as outlined in table 62, and are the higher of that derived using the RMBS credit criteria outlined above and that calculated as an additional minimum loss projection in accordance with this subsection.

Table 62

Rating Category Minimum Loss Projections For Concentrated Pools
Maximum potential rating	Minimum loss parameters
AAA	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'AAA' loss severity and the weighted-average loss severity at 'AAA' for the pool, plus the next 9 largest loss exposures at the 'AAA' loss severity
AA category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'AA' loss severity and the weighted-average loss severity at 'AA' for the pool, plus the next 7 largest loss exposures at the 'AA' loss severity
A category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'A' loss severity and the weighted-average loss severity at 'A' for the pool, plus the next 5 largest loss exposures at the 'A' loss severity
BBB category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'BBB' loss severity and the weighted-average loss severity at 'BBB' for the pool, plus the next 3 largest loss exposures at the 'BBB' loss severity
BB category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'BB' loss severity and the weighted-average loss severity at 'BB' for the pool, plus the next largest loss exposure at the 'BB' loss severity
B category	Highest balance loan liquidated at the greater of 50% loss severity, the loan's own 'B' loss severity and the weighted-average loss severity at 'B' for the pool

We may also adapt the above approach to assess the features (such as arrears status) of particular loans that account for the top X loans (X being the number of loans as determined in accordance with table 62), to capture any potential for loss on those loans to be greater than that determined under the calculation outlined in table 62.

REVISIONS AND UPDATES

This article was originally published on April 4, 2024.

Changes introduced after original publication:

• On May 17, 2024, we republished this criteria article to add a country-specific section for China, as China was added to the scope of our global methodology, on the same date, following our "Request For Comment: Global Methodology And Assumptions: Assessing Pools Of Residential Loans (China)," published Nov. 30, 2023. At the same time, we made nonmaterial changes to correct chart number references in the text, the display of the DTI curve for Korea), and the y-axis display in the LTV curve for Korea.
• On June 21, 2024, we republished this criteria article to add country-specific sections for Australia, New Zealand, and Singapore, as on the same date, these countries were added to the scope of our global methodology (with an effective date of June 28, 2024). This follows our "Request For Comment: Global Methodology And Assumptions: Assessing Pools Of Residential Loans (Australia, New Zealand, Singapore)," March 25, 2024. At the same time, we changed the document to present the country supplements in alphabetical order, made some other conforming changes, and updated the contact information.

RELATED PUBLICATIONS

For related publications, see the corresponding section in "Global Methodology And Assumptions: Assessing Pools Of Residential Loans," Jan. 25, 2019.

APAC Analytical Contacts:	Narelle Coneybeare, Sydney + 61 2 9255 9838; narelle.coneybeare@spglobal.com
	Calvin C Leong, Melbourne + 61 3 9631 2142; calvin.leong@spglobal.com
	Hiroshi Sonoda, Tokyo (81) 3-4550-8474; hiroshi.sonoda@spglobal.com
	Yalan Tao, Hong Kong + 852 2532 8033; yalan.tao@spglobal.com
Methodology Contacts:	Herve-Pierre P Flammier, Paris +33 1 44 20 73 38; herve-pierre.flammier@spglobal.com
	Andrew D Palmer, Melbourne + 61 3 9631 2052; andrew.palmer@spglobal.com
APAC Analytical Contacts:	Kate J Thomson, Melbourne + 61 3 9631 2104; kate.thomson@spglobal.com
	Yuji Hashimoto, Tokyo + 81 3 4550 8275; yuji.hashimoto@spglobal.com
	Jerry Fang, Hong Kong + 852 2533 3518; jerry.fang@spglobal.com
Methodology Contacts:	Andrew M Bowyer, CFA, London + 44 20 7176 3761; andrew.bowyer@spglobal.com