October 4, 2006
VIA FedEx and EDGAR
Mr. Jim B. Rosenberg
Senior Assistant Chief Accountant
Division of Corporation Finance
U.S. Securities and Exchange Commission
100 F Street, NE
Washington, DC 20549
Re: | Genworth Financial, Inc. |
Form 10-K for the fiscal year ended December 31, 2005
File No. 1-32195
Dear Mr. Rosenberg:
Reference is made to the letter dated August 25, 2006 to Michael D. Fraizer, Chairman, President and Chief Executive Officer, Genworth Financial, Inc. (Genworth or we), setting forth the Staffs comments on the above-referenced document (the Comment Letter). We are submitting this letter in response to the Staffs Comment Letter. The headings and numbered paragraphs in this letter correspond to the original headings and numbered paragraphs in the Comment Letter. For ease of reference, we have repeated the Staffs comments in bold text preceding each of our responses.
Managements Discussion and Analysis of Financial Condition and Results of Operations
Critical accounting policies
Deferred acquisition costs, page 117
1. | You periodically test DAC and PVFP for recoverability and increase amortization of DAC or benefit reserves when experience regarding expected expenses, investment returns, mortality, morbidity, withdrawals or lapses becomes unfavorable. Your disclosure should provide investors with a fuller understanding of the uncertainties in applying these critical accounting estimates and the likelihood that materially different amounts would be reported under different conditions or using different assumptions. Please provide in disclosure-type format an expanded discussion and quantification of the uncertainties in evaluating DAC and PVFP recoverability, the combined effect that such changes in estimate have had on your financial statements for each period presented and the effect that reasonably likely changes in the key assumptions underlying these estimates as of the latest balance sheet may have on your financial statements. |
Mr. Jim B. Rosenberg
Page 2
October 4, 2006
Please see Exhibit A to this letter for our proposed disclosure in future filings, which is marked to show changes from our 2005 Annual Report on Form 10-K.
Insurance liabilities and reserves, page 118
2. | Please provide an expanded description and quantification in disclosure-type format of the methodologies and assumptions used to determine your case and IBNR reserves for mortgage insurance. For example, this description might include a discussion of alternative models used, the strengths and weaknesses of each model and an explanation of why a specific model was ultimately chosen over the other models considered. Also, discuss judgmental techniques utilized when statistical data is insufficient or unavailable in estimating your reserves and your procedures for determining reserve adequacy at December 31, 2005. |
The IBNR reserves, case reserves and related loss adjustment expenses for our mortgage insurance business are based on our historical experience. Frequency and severity of loss factors are developed from this experience and are applied to current delinquencies reported to us by mortgage lenders and our estimate of unreported delinquencies. Factors considered in setting these reserves include historical trends in defaults, claim severity and frequency patterns, the category of delinquency (i.e., age and progression of delinquency) and loan coverage and unpaid principal balance. In addition, prevailing legal, economic, social and regulatory trends and their expected future developments are considered. There are also reserves established based on independent appraisals of the underlying properties and other available information at the time of the establishment of the reserves. The reserves are revised as updated information becomes available.
Our methods for estimating mortgage insurance IBNR and case reserves have been applied consistently, utilizing judgment in our loss models, where appropriate. In certain extraordinary circumstances, when statistical data is insufficient or unavailable, we will employ alternative estimation techniques that are more judgmental in nature. For example, as further described in our response to comment number 4 herein, we employed an alternative method for estimating losses incurred as a result of Hurricanes Katrina and Rita, developed based on our actual loss experience with Hurricane Andrew. This alternative method resulted in approximately $5 million lower case reserves as of December 31, 2005 than would have otherwise resulted from our normal reserving process.
We supplementally advise the Staff that as of December 31, 2005, our case and IBNR reserves were $290 million and $40 million, respectively, for our mortgage insurance business. If the reserves for our mortgage insurance business become material to our overall consolidated financial position, we will supplement our disclosures regarding these reserves.
Mr. Jim B. Rosenberg
Page 3
October 4, 2006
3. | Your accounting for the mortgage insurance business appears to depend significantly on data provided by loan servicers and other third parties. For example, under the delegated underwriting program, borrower default risk analysis is performed by lenders. Please describe in disclosure-type format how you provide for these data limitations in your accounting for mortgage insurance revenues and expenses, including how you compensate for factors you might otherwise evaluate in making an underwriting decision that are not considered by AUS. Include a description of your methods for validating the accuracy and completeness of data received from third parties. |
The accounting for elements of our mortgage insurance business does depend to a certain extent on data and other information provided to us by loan servicers and third-parties related to our delegated underwriting programs.
All data received and process decision determinations derived from within our delegated underwriting programs are based upon our own credit policy guidelines and derived from criteria that are based on our own pricing, compliance and risk standards set forth within our master policy agreements and delegated mortgage insurance arrangements. In addition, these agreements contain representations and warranties from respective third-parties with respect to the accuracy and completeness of certain data and information.
All data and information received through our delegated underwriting programs are subject to oversight by us including training and monitoring programs to ensure the data and information meet our guidelines and standards. In addition, the underlying processes are subject to underwriting and operating oversight procedures as well as regular procedural and process audits. We believe our combined oversight and audit processes provide reasonable assurance that financial and operational data received from third-parties is reliable for our accounting and financial reporting purposes in all material respects.
4. | You state that the 2005 impact associated with Hurricanes Katrina and Rita was based on preliminary estimates from available information and may be affected by factors, such as the pace of economic recovery in the affected areas, while 42% of your primary risk in-force as of December 31, 2005 was located in the Southeast and South Central U.S. regions. Please provide expanded discussion and quantification in disclosure-type format of the effect that reasonably likely changes in preliminary estimates as of the latest balance sheet may have on your financial statements. |
As discussed in Item 1- Business in our 2005 Annual Report on Form 10-K, our primary mortgage insurance covers default risk on first mortgage loans. Our insurance covers a specified coverage percentage of a claim amount consisting of unpaid loan principal, delinquent interest and certain expenses associated with the default and subsequent foreclosure. Our insurance does not cover property or casualty loss. The insured property must be undamaged or restored to its original condition (except for reasonable wear and tear) before we are obligated to pay a claim on the defaulted mortgage loan. Consequently, our exposure to catastrophes is substantially less than that of a property and casualty insurer.
Mr. Jim B. Rosenberg
Page 4
October 4, 2006
As of December 31, 2005, we had insured certificates on approximately 35,000 properties within the states that Federal Emergency Management Agency (FEMA) designated as federal disaster areas. While these states are in the Southeastern and South Central regions of the U.S., which as of December 31, 2005, represented 42% of our domestic primary risk in-force, the impacted states (i.e., Louisiana, Mississippi, Alabama and Texas) within those regions represent 11% of our domestic primary risk in-force.
Specific loss reserves were established for properties located within Severely Impacted Areas, as defined by FEMA, and for those loans within this area designated by FEMA as eligible for Individual Assistance programs because of the anticipation of the insurance recoverability by borrowers from their property casualty coverage. As of December 31, 2005, 1,296 properties, accounting for 0.8% of our domestic primary risk in-force, met these criteria for which we had insured certificates.
We supplementally advise the Staff that for the 1,296 properties in Severely Impacted Areas, we utilized a model, developed based on our actual loss experience with Hurricane Andrew, to estimate our exposure to Hurricanes Katrina and Rita related losses that factored in projections for key model drivers such as: (i) property condition/extent of property damage; (ii) borrower employment status; and (iii) determination of insurance recovery level, if any. As of December 31, 2005, the reserve for losses for properties within Severely Impacted Areas was $6 million. As of June 30, 2006, this reserve balance is $3 million, reflecting $3 million of favorable reserve development.
Going forward, for alternative estimation methods, we will disclose the reserve balance and, if material, differences from normal reserving methods, reasons why and key sensitivities.
Unearned premiums, page 118
5. | You state that unearned premiums are earned over time in accordance with the estimated expiration of risk and that such estimation is inherently judgmental. Given the long durations associated with mortgage insurance, your disclosure should provide investors with a fuller understanding of the uncertainties in applying these critical accounting estimates and the likelihood that materially different amounts would be reported under different conditions or using different assumptions. Please provide in disclosure-type format an expanded discussion and quantification of the uncertainties in estimating the expiration of risk under your mortgage insurance contracts, the effect that changes in such estimates have had on your financial statements for each period presented and the effect that reasonably likely changes in key assumptions underlying these estimates as of the latest balance sheet may have on your financial statements. |
Mr. Jim B. Rosenberg
Page 5
October 4, 2006
Please see Exhibit B to this letter for our proposed disclosure in future filings, which is marked to show changes from our 2005 Annual Report on Form 10-K.
* * * * *
We acknowledge the following:
| Genworth is responsible for the adequacy and accuracy of the disclosure in its filings; |
| staff comments or changes to disclosure in response to staff comments do not foreclose the Commission from taking any action with respect to Genworths filings; and |
| Genworth may not assert staff comments as a defense in any proceeding initiated by the Commission or any person under the federal securities laws of the United States. |
* * * * *
Should you have any questions regarding this matter, please contact Victor C. Moses at (804) 662-2411 or Scott R. Lindquist at (804) 662-2680.
Sincerely, | ||||
/s/ Victor C. Moses |
/s/ Scott R. Lindquist | |||
Victor C. Moses | Scott R. Lindquist | |||
Senior Vice President and Acting Chief Financial Officer | Vice President and Controller | |||
(Principal Financial Officer) | (Principal Accounting Officer) |
cc:
Don Abbott,
Senior Staff Accountant, U.S. Securities and Exchange Commission
Frank Wyman,
Staff Accountant, U.S. Securities and Exchange Commission
Michael D. Fraizer,
Chairman, President and Chief Executive Officer, Genworth Financial, Inc.
Exhibits as per above
Exhibit A
Genworth Financial, Inc. 2005 Annual Report on Form 10-K
Critical Accounting Policies, Pages 117-118
Deferred acquisition costs. Deferred acquisition costs, or DAC, represent costs which vary with and are primarily related to the sale and issuance of our insurance policies and investment contracts that are deferred and amortized over the estimated life of the related insurance policies. These costs include commissions in excess of ultimate renewal commissions, solicitation and printing costs, sales material and some support costs, such as underwriting and contract and policy issuance expenses. DAC is subsequently amortized to expense, over the lives of the underlying contracts, in relation to the anticipated recognition of premiums or gross profits.
The amortization of DAC for traditional long-duration insurance products (including guaranteed renewable term life, life-contingent structured settlements and immediate annuities and long-term care insurance) is determined as a level proportion of premium based on commonly accepted actuarial methods and reasonable assumptions, established when the contract or policy is issued, about mortality, morbidity, lapse rates, expenses, and future yield on related investments. U.S. GAAP requires that assumptions for these types of products not be modified (or unlocked) unless recoverability testing deems them to be inadequate or a repricing event occurs. Amortization is adjusted each period to reflect policy lapse or termination rates as compared to anticipated experience. Accordingly, we could experience accelerated amortization of DAC if policies terminate earlier than originally assumed.
Amortization of DAC for annuity contracts without significant mortality risk and investment and universal life
products is based on estimated expected gross profits and is adjusted as those estimates are revised. Expected gross profits are adjusted quarterly to reflect actual experience to-date or for the unlocking
of underlying key assumptions based on experience studies such as mortality, withdrawal or lapse rates, investment margin or maintenance expenses. The estimation of expected gross profits is subject to change given the inherent uncertainty as to the
underlying key assumptions employed and the long duration of our policy or contract liabilities. Changes in expected gross profits reflecting the unlocking of key assumptions could result in a material increase or decrease in the amortization of DAC
depending on the magnitude of the change in underlying assumptions. Significant factors that could result in a material increase or decrease in DAC amortization for these products include material changes in withdrawal or lapse rates, investment
spreads or mortality assumptions. For the years ended December 31, 2005, 2004 and 2003, key assumptions were unlocked in our Protection and Retirement Income and Investments segments to reflect our current expectation of future investment
spreads and mortality.
The amortization of DAC for mortgage insurance is based on expected gross margins. Expected gross margins, defined as premiums less losses, are set based on assumptions for future persistency and loss development of the business. These assumptions are updated quarterly for actual experience to-date or as our expectations of future experience are revised based on experience studies. Due to the inherent uncertainties in making assumptions about future events, materially different experience from expected results in persistency or loss development could result in a material increase or decrease to DAC amortization for this business. For the years ended December 31, 2005, 2004 and 2003, key assumptions were unlocked in our Mortgage Insurance segment to reflect our current expectation of future persistency.
i
Exhibit A
The following table sets forth the increase (decrease) on amortization of DAC related to unlocking of key assumptions for the periods indicated:
Years Ended December 31, | ||||||||||
(Amounts in millions) | 2005 | 2004 | 2003 | |||||||
Protection |
$ | 2 | $ | (4 | ) | $ | | |||
Retirement Income and Investments |
3 | 5 | 22 | |||||||
Mortgage Insurance |
7 | 8 | | |||||||
Total |
$ | 12 | $ | 9 | $ | 22 | ||||
The following discussion includes disclosure of variations from current estimates of
DAC amortization and recoverability due to reasonably likely adverse changes in certain key assumptions. The impact of reasonably likely adverse changes described below is estimated individually, without consideration for any correlation among key
assumptions or among lines of business. The estimated adverse variation in the amortization and recoverability of our DAC due to changes in key assumptions is an estimate of possible variation that may occur in the future,
likely over a period of several calendar years.
The DAC amortization methodology for our variable products (variable annuities and variable universal life insurance) includes a long-term equity market average appreciation assumption of 8.5%. When actual returns vary from the expected 8.5% we assume a reversion to this mean over a 3- to 5-year period, subject to the imposition of ceilings and floors. The assumed returns over this reversion period are limited to the 85th percentile of historical market performance. Variation in equity market returns that could be considered reasonably likely would not have a material effect on the amortization of DAC.
We regularly review all of these assumptions and periodically test DAC for recoverability DAC to determine
if it is recoverable from future income. For deposit products, if the current present value of estimated future gross profits is less than the unamortized DAC for a line of business, a charge to income is recorded for additional DAC amortization
or for increased benefit reserves. For other products, if the benefit reserves plus anticipated future premiums and interest earnings for a line of business are less than the current estimate of future benefits and expenses (including any
unamortized DAC), a charge to income is recorded for additional DAC amortization or for increased benefit reserves. Continued low interest rates and lower than expected termination rates on older issued long-term care insurance policies have
reduced our margins on this business. Additional adverse variations in morbidity, mortality, termination rates or interest rates that could be considered reasonably likely to occur in the future would result in our long-term care insurance DAC being
no longer fully recoverable. Variation that we consider reasonably likely would result in additional amortization of DAC of up to $11 million assuming there is no increase in future premiums. However, more adverse variation could result in
additional amortization of DAC or the establishment of additional benefit reserves, while any favorable variation would result in
ii
Exhibit A
additional margin in our DAC recoverability analysis and would result in higher earnings recognition over the remaining duration of the in-force block. We expect future sales of profitable long-term care insurance will improve margins, which would reduce the likelihood that adverse variation in key assumptions would result in our DAC being no longer fully recoverable. As of December 31, 2005, we had $1.6 billion of DAC related to our long-term care insurance business. We believe all of our other businesses have sufficient future income, where the related DAC would be recoverable under adverse variations in morbidity, mortality, withdrawal or lapse rate, maintenance expense or interest rates that could be considered reasonably likely to occur. For the years ended December 31, 2005, 2004 and 2003, there were no material charges to earnings as a result of our DAC recoverability testing.
Unfavorable experience with regard to expected expenses, investment returns, mortality, morbidity, withdrawals or lapses may cause us to increase the amortization of DAC or to record a charge to increase
benefit reserves. In recent years, the portion of estimated product margins required to amortize DAC and PVFP has increased in most lines of our business, with the most significant impact on investment products, primarily as the result of lower
investment returns.
Present value of future profits. In conjunction with the acquisition of a block of life insurance policies or investment contracts, a portion of the purchase price is assigned to the right to receive future gross profits arising from existing insurance and investment contracts. This intangible asset, called the present value of future profits, or PVFP, represents the actuarially estimated present value of future cash flows from the acquired policies. PVFP is amortized, net of accreted interest, in a manner similar to the amortization of DAC.
The following discussion includes disclosure of variations from current estimates of PVFP recoverability due to reasonably likely adverse changes in certain key assumptions. The impact of reasonably likely adverse changes described below is estimated individually, without consideration for any correlation among key assumptions or among lines of business. The estimated adverse variation in the recoverability of our PVFP due to changes in key assumptions is an estimate of possible variation that may occur in the future, likely over a period of several calendar years.
We regularly review our assumptions and periodically test PVFP for recoverability in a manner similar to our treatment of DAC. Continued low interest rates and lower than expected termination rates on long-term care insurance policies have reduced our margins for this acquired business. Additional adverse variations in morbidity, mortality, termination rates or interest rates that could be considered reasonably likely to occur in the future would result in our long-term care insurance PVFP being no longer fully recoverable. Variation that we consider reasonably likely would result in additional amortization of PVFP of up to $33 million assuming there is no increase in future premiums. However, more adverse variation could result in additional amortization of PVFP or the establishment of additional benefit reserves, while any favorable variation would result in additional margin in our PVFP recoverability analysis and would result in higher earnings recognition over the remaining duration of the in-force block. As of December 31, 2005, we had $0.1 billion of PVFP related to our long-term care insurance business. We believe all of our other
iii
Exhibit A
businesses have sufficient future income, where the related PVFP would be recoverable under adverse variations in morbidity, mortality, withdrawal or lapse rate, maintenance expense or interest rates that could be considered reasonably likely to occur. For the years ended December 31, 2005, 2004 and 2003, there were no material charges to earnings as a result of our PVFP recoverability testing.
iv
Exhibit B
Genworth Financial, Inc. 2005 Annual Report on Form 10-K
Critical Accounting Policies, Page 118
Unearned Premiums. For single premium insurance contracts, we recognize premiums over the policy life in accordance with the estimated expiration of risk. We
recognize a portion of the revenue in premiums earned in the current period, while the remaining portion is deferred as unearned premiums and earned over time in accordance with the estimated expiration of risk. If single premium policies are
canceled and the premium is non-refundable, then the remaining unearned premium related to each cancelled policy is recognized to earned premiums upon notification of the cancellation. Estimation of risk expiration on which we base premium
recognition is inherently judgmental and is based on actuarial analysis of historical experience. We periodically review our premium earnings recognition models with any adjustments to the estimates reflected in current period
earnings.
As of December 31, 2005, we had $3.6 billion of unearned premiums of which $1.8 billion related to our international mortgage insurance business. We recognize international mortgage insurance unearned premiums over a period of up to 25 years, most of which are recognized between 3 and 7 years from issue date. The recognition of earned premiums for our international mortgage insurance business involves significant estimates and assumptions as to future loss development and policy cancellations. These assumptions are based on our historical experience and our expectations of future performance, which are highly dependent on assumptions as to long-term macroeconomic conditions including interest rates, home price appreciation and the rate of unemployment. We periodically review our estimate of risk expiration and adjust this estimate based on actual experience and changes in our expectation of future performance with any adjustments reflected in current period earnings. For the years ended December 31, 2005, 2004 and 2003, increases to international mortgage insurance earned premiums as a result of adjustments made to our estimation of risk expiration were $21 million, $6 million and $0 million, respectively.
The following discussion includes disclosure of variations from current estimates of risk expiration due to reasonably likely changes in certain key assumptions. The estimated variation of our expiration of risk due to changes in key assumptions is an estimate of possible variation that may occur in the future, likely over a period of several calendar years.
Our estimate of risk expiration for our international mortgage insurance business is subject to change given the inherent uncertainty as to the underlying loss development and policy cancellation assumptions and the long duration of our international mortgage insurance policy contracts. Actual experience that is different than assumed for loss development or policy cancellations could result in a material increase or decrease in the recognition of earned premiums depending on the magnitude of the difference between actual and assumed experience. Loss development and policy cancellation variations that could be considered reasonably likely to occur in the future would result in accelerated or decelerated recognition of earned premiums that would result in an increase in net earnings of up to $40 million or a decrease in net earnings of up to $20 million, depending on the magnitude of variation experienced. It is important to note that the variation discussed above is not meant to be a best-case or worst-case scenario, and therefore, it is possible that future variation may exceed the amounts discussed above.
v
Exhibit B
The remaining portion of our unearned premiums relates to our payment protection insurance and long-term care insurance businesses where the underlying assumptions as to risk expiration are not subject to significant uncertainty. Accordingly, changes in underlying assumptions as to premium recognition we consider being reasonably likely for these businesses would not result in a material impact on net earnings.
vi