1. Standard & Poor's Ratings Services is requesting comments on its proposed changes to its methodology for rating mortgage insurers, specifically on the proposed key credit factors for the mortgage insurance (MI) industry. We are publishing this article to help market participants better understand these key credit factors.

2. This article should be read in conjunction with our insurance criteria (see "Insurers: Rating Methodology," published May 7, 2013, "Methodology: Mortgage Insurer Capital Adequacy" request for comment, [the "MI capital model"] published Aug. 12, 2014, and "Principles Of Credit Ratings," published Feb. 16, 2011). We intend to rate mortgage insurers consistent with the methodology in "Insurers: Rating Methodology" ("the framework"), except for the specific differences mentioned below.

3. If adopted, these proposed criteria would partly amend "Insurers: Rating Methodology" by expanding its scope as described below. For insurers already in scope with no significant mortgage (re)insurance business, there would be no change. The criteria constitute specific methodologies and assumptions under "Principles of Credit Ratings," published Feb. 16, 2011.

SCOPE OF THE PROPOSED CRITERIA

4. The proposed criteria would apply to all global-scale foreign and local currency, long-term issuer credit, financial strength, and financial enhancement ratings on mortgage insurers and insurers or insurance groups for which MI is a material part of the business.

SUMMARY OF THE PROPOSED CRITERIA

5. Under the proposal, we would rate mortgage insurers by applying Standard & Poor's insurance ratings framework (as well as the criteria in the Related Criteria section). We would assess MI in the same way as other insurers in the scope of the framework except for the following factors which, in our view, warrant specific methodologies for this sector:

• Insurance industry and country risk assessment (IICRA), reflecting the structural aspects of the mortgage and housing markets, the regulatory environment, the products provided, and the state of relevant macroeconomic factors;
• Competitive position, to take into account the monoline nature of the industry as well as the limited number of writers;
• Capital and earnings, which incorporates the factors discussed in the MI capital model request for comment article; and
• Liquidity.

SPECIFIC QUESTIONS FOR WHICH WE ARE SEEKING A RESPONSE

6. Standard & Poor's is seeking market feedback on its proposed criteria with respect to the following questions:

• Do the criteria incorporate the key factors affecting mortgage insurers' creditworthiness?
• Do you agree with the main variables for assessing the different factors?
• Are there other areas of the framework that should be adapted and included in this Key Credit Factors (KCF) article, or areas in the KCF that you view as unnecessary, as they would already be addressed in the framework?
• Do you agree with the way the proposed insurance industry and country risk assessment (IICRA) is reached and with how it could affect insurer ratings? If not, what alternatives would you propose?
• Are there any aspects of the proposal that are, in your view, ambiguous and require further clarification? Please specify the instances and paragraph numbers where you feel the proposal is ambiguous and elaborate on the nature of the ambiguity.

IMPACT ON OUTSTANDING RATINGS

7. If adopted as proposed, we expect the application of the proposed KCF criteria--combined with the proposed MI capital model criteria--to result in few, if any, rating changes. However, this assessment is based on historical or limited data used to test the proposed criteria, and our view of risk-adjusted capitalization for highly rated entities could become more negative with potential for downgrades.

RESPONSE DEADLINE

8. We encourage interested market participants to submit their written comments on the proposed criteria by Sept. 30, 2014 to http://www.standardandpoors.com/criteriaRFC/en/us. We will review and take such comments into consideration before publishing our definitive criteria once the comment period is over. Generally, Standard & Poor's Ratings Services (Ratings Services) may, in cases when the commenter has not requested confidentiality, publish comments in their entirety, except when the full text, in Ratings Services' view, would be unsuitable for reasons of tone or substance.

METHODOLOGY

Insurance Industry And Country Risk Assessment

9. The IICRA addresses the risks typically faced by insurers operating in specific industries and countries, and is generally determined at a country or regional level. The IICRA provides the context for our analysis of an insurer's business risk profile because industry and country risks are closely linked with the analysis of competitive position, as is the case for most corporate sectors. We assess nine IICRA subfactors in the analysis of insurance companies.

10. For mortgage insurers, we are retaining most of the IICRA criteria from the framework (see section VI.B2 of "Insurers: Rating Methodology") and are only proposing three specific differences for MI (see paragraph 31; Table 4: Alternative Metrics For Assessing Return on Equity [ROE], the fifth IICRA subfactor; and paragraphs 47-52, Product risk [the sixth IICRA subfactor] in the framework).

11. As part of this proposal, each MI sector in countries where we rate mortgage insurers would receive an IICRA assessment unless there was insufficient information. In this case, we would default to the property/casualty (P/C) IICRA assessment for that country. Paragraphs 12-14 below address the proposed specific methodologies where MI IICRA would be assessed differently than for other insurance sectors in the scope of the framework. (See table 3 in the Appendix for the indicative IICRA scores.)

MI-specific application of IICRA criteria

12. In the framework, we may assess a life, health, or P/C insurance market as less developed if premiums are less than 1.5% of GDP (see paragraph 31 of the framework). Less-developed markets have IICRA scores capped at "moderate." To assess an MI market as less developed, we would substitute premiums being less than 1.5% of GDP with the percentage of insured mortgages in the mortgage origination market, which is more reflective of the development of this particular market. This figure includes both private- and public-sector mortgage insurance. We expect markets where mortgage insurance is written on under 10% of the total originations to constitute a less-developed market, where the MI IICRA assessment is limited to "moderate."

ROE: Alternative metrics

13. The framework assesses the fifth subfactor, ROE, based on the ROE of at least 60% of industry participants by premiums (including all rated participants) if available, or else of all rated participants. This may be based on individual insurer data or aggregated data produced by regulators or industry associations. For MI, while we retain ROE as the primary metric, when ROE is not available we propose to use alternative metrics to inform our view of the sector's earnings (see table 1, which for MI would substitute table 4 of the framework).

Table 1

Alternative Metrics For Assessing Return On Equity
	1 (positive)*	3 (neutral)*	6 (negative)*
Return on revenue and combined ratio	We expect the average return on revenue to exceed approximately 60% and the average combined ratio to be approximately 60% or lower.	The assessment is neither positive nor negative.	We expect an average return on revenue of approximately 20% or lower or an average combined ratio that exceeds approximately 100%.
*Unless there is insufficient evidence to form an opinion or the available evidence suggests excessive risk-taking is taking place, in which case the assessment is "negative" (see paragraphs 44 and 45 of the framework).

Product risk

14. Some product-specific elements can cause ROE, or the metrics in Table 1, to be more volatile over time (see paragraph 47 of the framework for more detail). We may identify and assess further industry- or country-specific sources of volatility stemming from product risk. For mortgage insurers, these product risks are different from the product risks affecting P/C, life, and health insurers. Examples of product risks affecting MI include:

• Structural aspects of the mortgage market, including the structure of lending and funding mechanisms, legal regime, lender accountability, borrower recourse, and the tax deductibility of mortgage interest payments;
• Portfolio risks, including the type of loans, credit risk characteristics, and the quality of underwriting; and
• Macroeconomic risks relevant to MI, including employment levels and housing prices.

Competitive Position

15. Under the framework, we assess the level of an insurer's competitive position under six subfactors as "positive," "neutral," or "negative" (see table 5 of the framework):

• Operating performance,
• Differentiation of brand or reputation,
• Market position,
• The level of controlled distribution channels,
• Geographic diversification, and
• Other diversification.

16. For competitive position, to take into account the differences between the mortgage insurance sector and the broader insurance sector, we are proposing a specific application to MI regarding market position, geographic diversification, and other diversification. The remaining subfactors remain unchanged from the framework.

Market position

17. Market position is assessed primarily by an insurer's share of gross premiums for the market where it participates. When assessing market share, we propose to apply new insurance written (NIW) or the international equivalent instead of gross premiums. In addition, in Table 7 of the framework, we'd add mortgage insurance as an additional P/C line of business.

Geographic diversification

18. Diversification, particularly geographic, is fundamental to the insurance business. The subfactor is assessed as "positive," "neutral," or "negative" based on:

• The insurer's geographic presence, i.e., the insurer's footprint in the context of the surface area and size of markets where it writes business; and
• The level of insurance penetration, defined as the ratio of life and P/C premiums to GDP, in that geographic area.

19. We propose to define insurance penetration for MI as the percentage of insured mortgages in the mortgage origination market. We expect markets where mortgage insurance is written on greater than 10% of the total originations to constitute high penetration and all others to be low penetration.

Other diversification

20. As mortgage insurers are typically monoline companies, we expect to assess other diversification as "negative."

Capital And Earnings

21. Capital and earnings measures an insurer's ability to absorb losses by assessing capital adequacy prospectively, using quantitative and qualitative measures. Under the framework, an insurer's capital and earnings are assessed according to three subfactors:

• Regulatory capital adequacy,
• Capital adequacy, and
• Representativeness of modeling.

22. We are proposing an MI-specific application for the capital adequacy (see paragraphs 23-26) and representativeness of modeling (see paragraph 27) subfactors.

Capital adequacy

23. The capital adequacy subfactor considers capital prospectively by evaluating the amount of capital available relative to capital requirements that an insurer is likely to hold over the current and next two years to cover, over the expected life of its portfolio, losses from the various risks it carries. To assess capital adequacy for mortgage insurers (paragraph 97 of the framework), we propose to apply the framework except that an MI-specific capital model would assess the sources (rather than total adjusted capital [TAC]) and uses (rather than risk-based-capital [RBC] requirements) of capital (as detailed in the MI capital model request for comment). Also, for MI, Table 9 of the framework would be replaced with Table 2 below.

Table 2

Capital Adequacy Assessment For Mortgage Insurers
Score	Assessment	Guidance
1	Extremely strong	Prospective sources of capital exceed prospective uses at the 'AAA' stress level.
2	Very strong	Prospective sources of capital are below the prospective uses at the 'AAA' stress level but above, or only slightly below, the prospective uses at the 'AA' stress level.
3	Strong	Prospective sources of capital are slightly above the prospective uses at the 'A' stress level, but significantly below the prospective uses at the 'AA' stress level.
4	Moderately strong	Prospective sources of capital are significantly above the prospective uses at the 'BBB' stress level, but slightly below the prospective uses at the 'A' stress level.
5	Upper adequate	Prospective sources of capital are slightly above the prospective uses at the 'BBB' stress level, but significantly below the prospective uses at the 'A' stress level.
6	Lower adequate	Prospective sources of capital are at or near the prospective uses at the 'BBB' stress level.
7	Less than adequate	Prospective sources of capital are at or near prospective uses at the 'BB' stress level.
8	Weak	Prospective sources of capital are significantly below the prospective uses of capital at the ‘BB’ stress level.
Note: Rating symbols in this table refer to "stress levels" according to the MI capital model request for comment.

24. Our approach is different from the present value RBC approach, which addresses the multiline insurers we rate. For multiline insurers, forecasting sources and uses is extremely complex. Although the RBC model measures the impact of the stressed risk variables over the expected lives of the assets and liabilities, the volatility used to create the stressed scenarios is based on potential movements expected over one year. In other words, we are seeking to capture the present value of expected economic losses (change in shareholder equity/policyholder surplus) experienced over a year, to a degree of certainty that is commensurate with the rating. In addition, the RBC output of the multiline insurance model allows for an implicit line of business diversification credit--a feature not applicable to mortgage insurers.

25. Some of the scoring benchmarks proposed in Table 2 differ from those in the insurance framework, because of differences between the proposed MI capital model and the insurance capital model. The proposed MI capital model is based on a 10-year projection of capital sources and uses at various stress levels, where "sources" are primarily initial capital and premium and investment income, and uses are primarily insured losses (net of reinsurance recoveries) and general and administrative expenses. This differs from the insurance model, which is a point-in-time capital model that compares current capital with risk charges at the various stress levels, and projects capital and risk charges forward. Also, the proposed MI model is calibrated to include stresses at the 'AAA', 'AA', 'A', 'BBB', 'BB', and 'B' levels, whereas the insurance model does not calibrate below the 'BBB' stress level, using the magnitude of the deficiency to the 'BBB' stress as a benchmark for the "less than adequate" and "weak" scores.

26. The output of the RBC adequacy model for the multiline insurers is expressed as present value risk-adjusted capital being either redundant or deficient across targeted levels of risk-adjusted capitalization, consistent with the stress level. For MI, we analyze future values of sources and uses of funds rather than present values, thereby avoiding market perception of any weakness in the RBC, such as which discount rate to apply, while retaining a view of the ultimate capital needs of the entity under different stress levels as its insured portfolio runs off.

Representativeness of modeling

27. The representativeness of modeling subfactor determines whether the analysis of prospective capital adequacy has overstated or understated capital and earnings. With respect to paragraphs 117 and 118 of the framework, we propose not to apply to MI the "moderately negative" cap for companies below $1 billion TAC and the "negative" cap for companies below $250 million TAC. These are not relevant for MI, as the MI capital model proposal is consistent with RMBS criteria, which addresses much smaller and less diversified pools of assets.

Liquidity

28. Under our framework, the liquidity analysis centers on an insurer's ability to cover its liquidity needs, both on an ongoing basis and in moderately stressful market and economic conditions. The analysis is absolute, rather than relative to peers. An insurer's liquidity assessment results from that of four subfactors:

• Coverage of the insurer's confidence-sensitive liabilities;
• The possibility that the insurer would need to post collateral;
• The implications of covenants and ratings triggers in the insurer's financial arrangements; and
• The insurer's liquidity ratio.

29. In assessing the liquidity ratio for mortgage insurers, we also consider in the numerator premiums collected and investment income expected over the next 12 months. The denominator includes no net reserve charge, net catastrophe charge, or net premium charge. The net reserve charge contemplates circumstances wherein adverse deviation could result in an immediate claim payment, circumstances which we have not found to have had a significant impact on short-term liquidity needs within this sector. As well, MI is not subject to catastrophic events that would cause a need for more immediate claim payments, obviating the need for a catastrophe charge. With regards to the net premium charge, in an environment of heightened claims, new business is assumed to decline significantly (a substantial portion of the premiums written and received would be from the book currently in force at the beginning of the stress period). Instead, we propose the following liquidity ratio for MI:

30. The risk charges for the stressed liquid assets in the numerator are as per paragraph 97 of the framework.

Glossary

Housing price index

An index measuring the price changes of residential housing from period to period.

Loan-to-value (LTV/LVR) ratio

The ratio of the mortgage loan amount to the lesser of the property's appraised value or sale price.

Mortgage insurance

Mortgage insurance is an insurance policy that protects mortgage lenders from borrower defaults. The MI policy coverage can be purchased for a percentage of the loan value and covered costs or for the full amount. In the U.S., MI policy coverage on a loan is typically provided within a range of 20%-35% while in Canada and Australia MI covers 100% of the eligible claims.

New insurance written (NIW)

Original loan amounts on policies written during the current reporting period.

Loan type

Loan types refer to the specification of a loan, including the loan term, requirements, and amortization.

Unemployment rate

Percent of labor force that is unemployed.

APPENDIX

Related Criteria

• Methodology: Mortgage Insurer Capital Adequacy, Aug. 12, 2014
• Ratings Above The Sovereign: Corporate And Government Ratings--Methodology And Assumptions, Nov. 19, 2013
• Group Rating Methodology, Nov. 19, 2013
• Insurers Rating Methodology, May 7, 2013
• Enterprise Risk Management, May 7, 2013
• Methodology For Linking Short-Term And Long-Term Ratings For Corporate, Insurance, And Sovereign Issuers, May 7, 2013
• Methodology: Management And Governance Credit Factors For Corporate Entities And Insurers, Nov. 13, 2012
• Criteria For Assigning 'CCC+', 'CCC', 'CCC-', And 'CC' Ratings, Oct. 1, 2012
• Refined Methodology And Assumptions For Analyzing Insurer Capital Adequacy Using The Risk-Based Insurance Capital Model, June 7, 2010
• Principles Of Credit Ratings, Feb. 16, 2011

Indicative MI IICRA Scores

This section (including table 3) is not part of the proposed criteria but is provided for informational purposes. The following table provides indicative IICRA scoring for countries where we have interactively rated mortgage insurers with ratings that incorporate stand-alone analysis of mortgage insurance operations.

Table 3

Indicative MI IICRA Scores
Country	IICRA score	Country risk	Industry risk
U.S.	Intermediate risk	Very low risk	Moderate risk
Australia	Low risk	Very low risk	Low risk
Canada	Low risk	Very low risk	Low risk
Hong Kong	Low risk	Very low risk	Intermediate risk

Primary Credit Analysts:	Hardeep S Manku, Toronto 416-507-2547; hardeep.manku@standardandpoors.com
	Michael J Vine, Melbourne (61) 3-9631-2102; michael.vine@standardandpoors.com
	Marc Cohen, CFA, New York (1) 212-438-2031; marc.cohen@standardandpoors.com
Secondary Contacts:	Kevin T Ahern, New York (1) 212-438-7160; kevin.ahern@standardandpoors.com
	Farooq Omer, CFA, Hightstown (1) 212-438-1129; farooq.omer@standardandpoors.com
	Carmi Margalit, CFA, New York (1) 212-438-2281; carmi.margalit@standardandpoors.com
	Mark A Legge, Melbourne (61) 3-9631-2041; mark.legge@standardandpoors.com
	Terry Sham, CFA, FRM, Hong Kong (852) 2533-3590; terry.sham@standardandpoors.com
Criteria Officers:	Emmanuel Dubois-Pelerin, Global Criteria Officer, Financial Services, Paris (33) 1-4420-6673; emmanuel.dubois-pelerin@standardandpoors.com
	Lucy A Collett, Chief Credit Officer, Americas, New York (1) 212-438-6627; lucy.collett@standardandpoors.com
	Michelle M Brennan, European Financial Services Criteria Officer, London (44) 20-7176-7205; michelle.brennan@standardandpoors.com