Note 2 - Significant Accounting Policies Level 2 (Policies)

12 Months Ended

Dec. 31, 2013

Significant Accounting Policies [Abstract]

Our consolidated financial statements are prepared in accordance with GAAP and include the accounts of all wholly-owned subsidiaries. Companies in which we, or one of our subsidiaries, exercise significant influence (generally ownership interests ranging from 20% to 50%), are accounted for in accordance with the equity method of accounting. VIEs for which we are the primary beneficiary are consolidated, as described in Note 6. All intercompany accounts and transactions, and intercompany profits and losses, have been eliminated.

Use of Estimates

The preparation of financial statements in conformity with GAAP requires us to make estimates and assumptions that affect the reported amounts of assets and liabilities, the disclosure of contingent assets and liabilities at the dates of the financial statements and the reported amounts of revenues and expenses during the reporting periods. While the amounts included in our consolidated financial statements include our best estimates and assumptions, actual results may vary materially.

Liability Reserve Estimate, Policy [Policy Text Block]

Reserve for Losses and LAE

We establish reserves to provide for losses and LAE and the estimated costs of settling claims in both our mortgage insurance and financial guaranty segments in accordance with the accounting standard regarding accounting and reporting by insurance enterprises. Although this standard specifically excludes mortgage insurance from its guidance relating to the reserve for losses, we establish reserves for mortgage insurance as described below, using the guidance contained in this standard supplemented with other accounting guidance, due to the lack of specific guidance for mortgage insurance.

Estimating the loss reserves in both our mortgage insurance and financial guaranty business segments involves significant reliance upon assumptions and estimates with regard to the likelihood, magnitude and timing of each potential loss. The models, assumptions and estimates we use to establish loss reserves may prove to be inaccurate, especially during an extended economic downturn or a period of extreme market volatility and uncertainty. As such, we cannot be certain that our reserve estimate will be adequate to cover ultimate losses on incurred defaults.

Commutations, recaptures and other negotiated terminations of our insured risks in both our mortgage insurance and financial guaranty segments provide us with an opportunity to exit exposures for an agreed upon payment, or payments, sometimes at an amount less than the previously estimated ultimate liability. Once all exposures relating to such policies are extinguished, all reserves for losses and LAE and other balances relating to the insured or reinsured policy generally are eliminated. Upon completion of a commutation, recapture or other negotiated termination, all such related balances, including deferred policy acquisition costs and unearned premiums, are generally reversed, with any remaining net gain or loss typically recorded through provision for losses. We take into consideration the specific contractual and economic terms for each individual agreement when accounting for our commutations, recaptures or other negotiated terminations, which may result in differences in the accounting for these transactions.

Mortgage Insurance

In the mortgage insurance segment, the default and claim cycle begins with the receipt of a default notice from the servicer. Reserves for losses are established upon receipt of notification by servicers that a borrower has missed two monthly payments, which is when we consider a loan to be in default for financial statement and internal tracking purposes. We also establish reserves for associated LAE, consisting of the estimated cost of the claims administration process, including legal and other fees and expenses associated with administering the claims process. We maintain an extensive database of claim payment history and use models based on a variety of loan characteristics, including the status of the loan as reported by its servicer and the type of loan product, to determine the likelihood that a default will reach claim status. Our process includes forecasting the impact of our loss mitigation efforts in protecting us against fraud, underwriting negligence, breach of representation and warranties, inadequate documentation of submitted claims and other items that may give rise to insurance rescissions or cancellations and claim denials, to help determine the rate at which loans are expected to move to claim (“default to claim rate”). Lastly, we project the amount that we will pay if a default becomes a claim (referred to as “claim severity”), which is also impacted by loss mitigation activity associated with claim curtailments due to servicer noncompliance with our insurance policies and servicing guidelines. When there is a claim under primary mortgage insurance, the coverage percentage is applied to the claim amount, which consists of the unpaid loan principal, plus past due interest (for which our liability is contractually capped at a maximum of two years) and certain expenses associated with the default, to determine our maximum liability. Based on these estimates, we arrive at our estimate of loss reserves as of that time.

With respect to loans that are in default, considerable judgment is exercised as to the adequacy of reserve levels. Loss reserves are increased as defaulted loans age, up to a maximum of 240 days in default because historically, as defaulted loans age, they have been more likely to result in foreclosure, and therefore, have been more likely to result in a claim payment. In the past, as the default proceeded towards foreclosure, there was generally more certainty regarding these estimates. However, in the current environment in which many foreclosures have been delayed or not yet pursued, significant uncertainty remains with respect to the ultimate resolution of aged defaults. This uncertainty requires management to use considerable judgment in estimating the rate at which these loans will result in claims. If a default cures, the reserve for that loan is removed from the reserve for losses and LAE. Once a claim is submitted, reserves are further increased to reflect the fact that the default has moved closer to resulting in a claim payment.

We also establish reserves for defaults that we estimate have been incurred but have not been reported (“IBNR”) to us on a timely basis by the servicer as well as for previously rescinded policies and denied claims that we estimate will be reinstated and subsequently paid. We generally give the insured up to 90 days to challenge our decision to rescind coverage before we consider a policy to be rescinded and remove it from our defaulted inventory; therefore, we currently expect only a limited percentage of policies that were rescinded to be reinstated. We currently expect a significant percentage of claims that were denied to be resubmitted as a perfected claim and ultimately paid. Most often, a claim denial is the result of a servicer’s inability to provide the loan origination file or other servicing documents for review. Under the terms of our master insurance policy with our lending customers, our insureds have up to one year after the acquisition of borrower’s title to provide to us the necessary documents to perfect a claim. All estimates are periodically reviewed and adjustments are made as they become necessary.

We do not establish reserves for loans that are in default if we believe that we will not be liable for the payment of a claim with respect to that default. For example, for those defaults in which we are in a “second loss position” (i.e., we are not required to make a payment until a certain aggregate amount of losses have already been recognized on a given set of loans), we initially calculate the reserve for defaulted loans in the transaction as if there were no deductible. If the existing deductible for a given structured transaction is greater than the aggregate reserve amount for the defaults contained within the transaction, we do not establish a reserve for the defaults, or if appropriate, we record a partial reserve. We do not establish loss reserves for expected future claims on insured mortgages that are not in default. See “—Reserve for Premium Deficiency (“PDR”)” below for an exception to this general principle.

For purposes of reserve modeling, loans are aggregated into groups using a variety of factors. The attributes used to define the groups include, but are not limited to, the default status of the loans (i.e., number of days in default), product type (i.e., Prime, Alternative-A (“Alt-A”) or Subprime), type of insurance (i.e., primary or pool), policy origination year, loss position (i.e., with or without a deductible) and the state where the property is located (segregated into three state groups in order to adjust for differences in foreclosure timing). We use an actuarial projection methodology referred to as a “roll rate” analysis that uses historical claim frequency information to determine the projected ultimate default to claim rates for each product and default status. The default to claim rate also includes our estimates with respect to expected insurance rescissions and claim denials, which have the effect of reducing our default to claim rates. Since 2009, we have experienced an elevated level of insurance rescissions and claim denials for various reasons, including, without limitation, underwriting negligence, fraudulent applications and appraisals, breach of representations and warranties and inadequate documentation, primarily related to our legacy portfolio. We expect our rescission and denial rates to remain at elevated levels as long as defaults related to our legacy portfolio represent a significant percentage of our total default portfolio. After estimating the default to claim rate, we estimate the severity of each product type, type of insurance and state grouping based on the average of recently observed severity rates. These average severity estimates are then applied to individual loan coverage amounts to determine reserves. Senior management regularly reviews the modeled frequency, rescission, denial and severity estimates, which are based on historical trends as described. If recent emerging or projected trends differ significantly from the historical trends used to develop the modeled estimates, management evaluates these trends in determining how they should be considered in its reserve estimates.

In addition, as part of our claims review process, we assess whether defaulted loans were serviced appropriately in accordance with our insurance policies and servicing guidelines. To the extent a servicer has failed to satisfy its servicing obligations, our policies provide that we may curtail the claim payment for such default, and in some circumstances, cancel coverage or deny the claim. Since 2011, claim curtailments have increased both in frequency and in size, which has contributed to a reduction in the severity of our claim payments during this period. While we cannot give assurance regarding the extent or level at which such claim curtailments will continue, we expect this trend to continue in light of well publicized issues in the servicing industry and our existing legacy portfolio of aged defaults.

The elevated levels of our loss mitigation actions (the rate of rescissions and denials since 2009 and claim curtailments more recently) have led to an increased risk of litigation. Under our master insurance policy, any suit or action arising from any right of the insured under the policy must be commenced within two years after such right first arose and within three years for certain other policies, including certain pool insurance policies. Recently, we have faced an increasing number of challenges from certain lender customers regarding our loss mitigation actions, which have resulted in some reversals of our decisions regarding rescissions, denials and curtailments. Although we believe that our loss mitigation actions are justified under our policies, if we are not successful in defending these actions in any potential legal or other actions, including negotiated settlements, we may need to reassume the risk on, and increase loss reserves for, those policies or pay additional claims. The assumptions embedded in our estimated default to claim rate on our in-force default inventory include an adjustment to our estimated rescission and denial rate, to account for the fact that we expect a certain number of policies to be reinstated and ultimately to be paid, as a result of valid challenges by such policy holders. As discussed above, we also establish reserves for IBNR defaults related to previously rescinded policies and denied or curtailed claims, which we believe are likely to be reinstated (in the case of previously rescinded policies) or resubmitted and paid (in the case of previously denied claims).

Financial Guaranty

In our financial guaranty segment, we recognize a net claim liability on our non-derivative transactions prior to an event of default (insured event) when there is evidence that credit deterioration has occurred for a particular policy and that the present value of the expected claim loss exceeds the unearned premium revenue. The expected claim loss is based on the probability-weighted present value of expected net cash outflows to be paid under, or in connection with, the policy. In measuring the net claim liability, we develop the present value of expected net cash outflows by using our own assumptions about the likelihood of various possible outcomes, including potential settlements or commutations, based on information currently available. We determine the existence of credit deterioration on directly insured policies based on periodic reporting from the insured party, indenture trustee or servicer and based on our surveillance efforts. These expected cash outflows are discounted using a risk-free rate. Our assumptions about the likelihood of outcomes, expected cash outflows and the appropriate risk-free rate are updated each reporting period. For assumed policies, we primarily use information provided by the ceding company, but also consider our specific knowledge of the credit for determining expected loss.

The risk management function in our financial guaranty business is responsible for the identification, analysis, measurement and surveillance of credit, market, legal and operational risk associated with our financial guaranty insurance contracts. Risk management is also primarily responsible for claims prevention and loss mitigation strategies. This discipline is applied during the ongoing monitoring and surveillance of each exposure in the portfolio. See Note 10 for further information.

Reserve For Premium Deficiency

Reserve for Premium Deficiency (“PDR”)

Insurance enterprises are required to establish a PDR if the net present value of the expected future losses and expenses for a particular product line exceeds the net present value of expected future premiums and existing reserves for that product line. We reassess our expectations for premiums, losses and expenses for our financial guaranty and mortgage insurance businesses at least quarterly and update our premium deficiency analysis accordingly. Expected future expenses include consideration of maintenance costs associated with maintaining records relating to insurance contracts and with the processing of premium collections. We also consider investment income in the premium deficiency calculation through the use of our pre-tax investment yield to discount certain cash flows for this analysis.

For our mortgage insurance business, we group our mortgage insurance products into two categories: first-lien and second-lien. To assess the need for a PDR on our first-lien insurance portfolio, we develop loss projections based on modeled loan defaults related to our current RIF. This projection is based on recent trends in default experience, severity and rates of defaulted loans moving to claim (such default to claim rates are net of our estimates of rescissions and denials), as well as recent trends in the rate at which loans are prepaid.

For our second-lien insurance business, we project future premiums and losses for this business using historical results to help determine future performance for both prepayments and claims. An estimated expense factor is then applied, and the result is discounted using a rate of return that approximates our pre-tax investment yield. This net present value, less any existing reserves, is recorded as a premium deficiency and the reserve is updated at least quarterly based on actual results for that quarter, along with updated transaction level projections.

For our financial guaranty business, to determine whether a premium deficiency charge is necessary, we compare projected earned premiums and investment income to projected future losses, LAE, unamortized deferred acquisition costs and maintenance costs. If the sum of the costs exceeds the amount of the revenues, the excess is first charged against deferred acquisition costs and is referred to as a premium deficiency charge.

Derivative instruments

Derivative Instruments

We provide a significant portion of our credit protection within our financial guaranty segment in the form of CDS, which are accounted for as derivatives. Derivative instruments are recorded at fair value and changes in fair value are recorded as such in the statement of operations. All of our derivative instruments are recognized in our consolidated balance sheets as either derivative assets or derivative liabilities. In many of our CDS transactions, primarily our corporate CDOs, we are required to make payments to our counterparty above a specified level of subordination, upon the occurrence of credit events related to the borrowings or bankruptcy of obligors contained within pools of corporate obligations or, in the case of pools of mortgage or other asset-backed obligations, upon the occurrence of credit events related to the specific obligations in the pool. When we provide a CDS as credit protection on a specific obligation, we generally guarantee the full and timely payment of principal and interest when due on such obligation. These derivatives have various maturity dates, but the majority of the net par outstanding of our remaining CDS transactions, including all of our corporate CDOs, mature within four years. See Note 4 for further information.

We recorded premiums and origination costs related to our CDS and certain other derivative contracts in change in fair value of derivative instruments and policy acquisition costs, respectively, on our consolidated statements of operations. Our classification of these contracts is the same whether we are a direct insurer or we reinsure these contracts.

Variable Interest Entities

VIEs

As a provider of credit enhancement, we have entered into insurance contracts with VIEs and derivative contracts with counterparties where we have provided credit protection directly on variable interests and, in some cases, obtained the contractual rights of our counterparties with respect to the VIEs. VIEs include corporations, trusts or partnerships in which equity investors do not have a controlling financial interest or do not have sufficient equity at risk to finance activities without additional subordinated financial support.

An entity is considered the primary beneficiary and is required to consolidate a VIE if its variable interest: (i) gives it the power to most significantly impact the economic performance of the VIE; and (ii) has the obligation to absorb losses or the right to receive residual benefits that could potentially be significant to the VIE. For all VIEs in which we have a variable interest, we assess whether we are the primary beneficiary. In determining whether we are the primary beneficiary, a number of factors are considered, including the structure of the entity, provisions in our insurance contracts that grant us additional rights to influence or control the economic performance of the VIE upon the occurrence of an event of default, a servicer termination event or the breach of a performance trigger, and our obligation to absorb significant losses. The breach of these performance tests or other events that give rise to our right to influence or control the economic performance of the VIE may occur, particularly if credit performance deteriorates. When we acquire control rights, we perform an analysis to reassess our involvement with these VIEs to determine whether we have become the primary beneficiary.

Fair Value of Financial Instruments, Policy [Policy Text Block]

Fair Value of Financial Instruments

Our estimated fair value measurements are intended to reflect the assumptions market participants would use in pricing an asset or liability based on the best information available. Assumptions include the risks inherent in a particular valuation technique (such as a pricing model) and the risks inherent in the inputs to the model. Changes in economic conditions and capital market conditions, including but not limited to, credit spread changes, benchmark interest rate changes, market volatility and changes in the value of underlying collateral or of any third-party guaranty or insurance, could cause actual results to differ materially from our estimated fair value measurements. We define fair value as the current amount that would be received to sell an asset or paid to transfer a liability in an orderly transaction between market participants at the measurement date. In the event that our investments or derivative contracts were sold, commuted, terminated or settled with a counterparty or transferred in a forced liquidation, the amounts received or paid may be materially different from those determined in accordance with the accounting standard regarding fair value measurements. Differences may also arise between our recorded fair value and the settlement or termination value with a counterparty based upon consideration of information that may not be available to another market participant. Those differences, which may be material, are recorded as realized gains/(losses) in our consolidated statements of operations in the period in which the transaction occurs.

In May 2011, the Financial Accounting Standards Board (“FASB”) issued an update to the accounting standard regarding fair value measurements and disclosure. This update clarifies the FASB’s intent about the application of existing fair value measurement and disclosure requirements and, in limited situations, changes certain principles or requirements for measuring fair value and disclosing information about fair value measurements. The amendments in this update do not require additional fair value measurements and are not intended to establish valuation standards or affect valuation practices outside of financial reporting. We adopted this update effective January 1, 2012. The adoption of this update did not have a significant impact on our fair value measurements. Additional disclosures regarding unobservable market inputs related to our Level III instruments required under this update are presented in Note 5.

When determining the fair value of our liabilities, we are required to incorporate into the fair value of those liabilities an adjustment that reflects our own non-performance risk. Our CDS spread is an observable quantitative measure of our non-performance risk and is used by typical market participants to determine the likelihood of our default. Assuming all other factors are held constant, as our CDS spread tightens or widens, it has the effect of increasing or decreasing, respectively, the fair value of our liabilities with a corresponding impact on our results of operations.

In accordance with GAAP, we established a three-level valuation hierarchy for disclosure of fair value measurements based on the transparency of inputs to the valuation of an asset or liability as of the measurement date. The hierarchy gives the highest priority to unadjusted quoted prices in active markets for identical assets or liabilities (Level I measurements) and the lowest priority to unobservable inputs (Level III measurements). The level in the fair value hierarchy within which the fair value measurement falls is determined based on the lowest level input that is significant to the measurement in its entirety. The three levels of the fair value hierarchy are defined below:


Level I	— Unadjusted quoted prices for identical assets or liabilities in active markets that are accessible at the measurement date for identical, unrestricted assets or liabilities;


Level II	— Prices or valuations based on observable inputs other than quoted prices in active markets for identical assets and liabilities; and


Level III	— Prices or valuations that require inputs that are both significant to the fair value measurement and unobservable. Level III inputs are used to measure fair value only to the extent that observable inputs are not available.

For markets in which inputs are not observable or are limited, we use significant judgment and assumptions that a typical market participant would use to evaluate the market price of an asset or liability. Given the level of judgment necessary, another market participant may derive a materially different estimate of fair value. These assets and liabilities are classified in Level III of our fair value hierarchy.

Available for sale securities, trading securities, VIE debt, derivative instruments and certain other assets are recorded at fair value as described in Note 5. All derivative instruments and contracts are recognized in our consolidated balance sheets as either derivative assets or derivative liabilities. All changes in fair value of trading securities, VIE debt, derivative instruments and certain other assets are included in our consolidated statements of operations. All changes in the fair value of available for sale securities are recorded in accumulated other comprehensive income (loss).

NIMS Derivative Assets and NIMS VIE Debt

NIMS derivative assets primarily represent derivative assets related to NIMS trusts that we are required to consolidate. NIMS VIE debt represents the debt of consolidated NIMS trusts, which we account for at fair value. The estimated fair value amounts of these financial instruments are derived from internally-generated discounted cash flow models. We estimate losses in each securitization underlying NIMS derivative assets or NIMS VIE debt by applying expected default rates separately to loans that are delinquent and those that are paying currently. These default rates are based on historical experience of similar transactions. We then estimate the rate of prepayments on the underlying collateral in each securitization, incorporating historical prepayment experience. The estimated loss and rate of prepayments are used to estimate the cash flows for each underlying securitization and NIMS bond, and ultimately, to produce the projected credit losses for each NIMS bond. The NIMS derivative assets and NIMS VIE debt are all categorized in Level III of the fair value hierarchy.

The following are descriptions of our valuation methodologies for financial assets and liabilities measured at fair value.

Investments

We are responsible for the determination of the value of all investments carried at fair value and the supporting methodologies and assumptions. To assist us in this responsibility, we utilize independent third-party valuation service providers to gather, analyze and interpret market information and estimate fair values based upon relevant methodologies and assumptions for various asset classes and individual securities. We perform monthly quantitative and qualitative analysis on the prices received from third parties to determine whether the prices are reasonable estimates of fair value. Our analysis includes: (i) a review of the methodology used by third-party pricing services; (ii) a comparison of pricing services’ valuations to other independent sources; (iii) a review of month-to-month price fluctuations; and (iv) a comparison of actual purchase and sale transactions with valuations received from third parties. These processes are designed to ensure that our investment values are accurately recorded, that the data inputs and valuation techniques utilized are appropriate and consistently applied and that the assumptions are reasonable and consistent with the objective of determining fair value.

U.S. government and agency securities—The fair value of U.S. government and agency securities is estimated using observed market transactions, including broker-dealer quotes and actual trade activity as a basis for valuation. U.S. government and agency securities are categorized in either Level I or Level II of the fair value hierarchy.

State and municipal obligations—The fair value of state and municipal obligations is estimated using recent transaction activity, including market and market-like observations. Evaluation models are used, which incorporate bond structure, yield curve, credit spreads and other factors. These securities are generally categorized in Level II of the fair value hierarchy or in Level III when market-based transaction activity is unavailable.

Money market instruments—The fair value of money market instruments is based on daily prices, which are published and available to all potential investors and market participants. As such, these securities are categorized in Level I of the fair value hierarchy.

Corporate bonds and notes—The fair value of corporate bonds and notes is estimated using recent transaction activity, including market and market-like observations. Spread models are used that incorporate issuer and structure characteristics, such as credit risk and early redemption features, where applicable. These securities are generally categorized in Level II of the fair value hierarchy or in Level III when market-based transaction activity is unavailable.

RMBS, CMBS, and Other ABS—The fair value of these instruments is estimated based on prices of comparable securities and spreads and observable prepayment speeds. These securities are generally categorized in Level II of the fair value hierarchy or in Level III when market-based transaction activity is unavailable. The fair value of any Level III securities is generally estimated by discounting estimated future cash flows.

Foreign government and agency securities—The fair value of foreign government and agency securities is estimated using observed market yields used to create a maturity curve and observed credit spreads from market makers and broker-dealers. These securities are categorized in Level II of the fair value hierarchy.

Hybrid securities—These instruments are convertible securities. The estimated fair value is derived, in part, by utilizing dealer quotes and observed bond and stock prices. For certain securities, the underlying security price may be adjusted to account for observable changes in the conversion and investment value from the time the quote was obtained. These securities are categorized in Level II of the fair value hierarchy.

Equity securities—The fair value of these securities is generally estimated using observable market data in active markets or bid prices from market makers and broker-dealers. Generally, these securities are categorized in Level I or II of the fair value hierarchy, as observable market data are readily available. A small number of our equity securities, however, are categorized in Level III of the fair value hierarchy due to a lack of market-based transaction data or the use of model-based evaluations.

Other investments—These securities primarily consist of deposit investments and short-term CDs, which are categorized in Level II and Level III of the fair value hierarchy, and a guaranteed investment contract held by one of our consolidated VIEs, which is categorized in Level III of the fair value hierarchy. The fair value of the Level III securities is generally estimated by discounting estimated future cash flows.

Derivative Instruments and Related VIE Assets/Liabilities

We define fair value as the current amount that would be exchanged to sell an asset or transfer a liability, other than in a forced liquidation. In determining an exit market, we consider the fact that most of our derivative contracts are unconditional and irrevocable and contractually prohibit us from transferring them to other capital market participants. Accordingly, there is no principal market for such highly structured insured credit derivatives. In the absence of a principal market, we value these insured credit derivatives in a hypothetical market where market participants include other monoline mortgage and financial guaranty insurers with similar credit quality to us, as if the risk of loss on these contracts could be transferred to these other mortgage and financial guaranty insurance and reinsurance companies. We believe that in the absence of a principal market, this hypothetical market provides the most relevant information with respect to fair value estimates.

We determine the fair value of our derivative instruments primarily using internally-generated models. We utilize market observable inputs, such as credit spreads on similar products, whenever they are available. When one of our transactions develops characteristics that are inconsistent with the characteristics of transactions that underlie the relevant market-based index that we use in our credit spread valuation approach, and more relevant inputs or projections become available and would represent the view of a typical market participant, we change to an approach that is based on that more relevant available information. This change in approach is generally prompted when the credit component, and not market factors, becomes the dominant driver of the estimated fair value for a particular transaction. There is a high degree of uncertainty about our fair value estimates since our contracts are not traded or exchanged, which makes external validation and corroboration of our estimates difficult, particularly given the current market environment, in which very few, if any, contracts are being traded or originated.

Our derivative liabilities valuation methodology incorporates our own non-performance risk by including our observable CDS spread as an input into the determination of the fair value of our derivative liabilities. Considerable judgment is required to interpret market data to develop the estimates of fair value. Accordingly, the estimates may not be indicative of amounts we could realize in a current market exchange or negotiated termination. Our derivative liability valuation is not counterparty specific and is intended to estimate the average exchange price between typical participants. The use of different market assumptions or estimation methodologies may have a material effect on the estimated fair value amounts or negotiated terminations. In a negotiated termination, certain factors unique to the counterparty may have a greater impact on the amount exchanged than in an estimated fair value amount between typical market participants and another market participant could have materially different views given the level of judgment associated with the valuation.

The significant unobservable inputs in the fair value measurement of our investment securities noted in the tables above include an interest rate used to discount the projected cash flows and an expected loss assumption. This expected loss assumption generally represents the principal shortfall we believe that a typical market participant would expect on our security as a result of the obligor’s failure to pay. In addition, our other investments include a guaranteed investment contract for which the counterparty’s non-performance risk is considered in the discount rate. Significant increases (decreases) in either the discount rates or loss estimates in isolation would result in a lower (higher) fair value measurement. Changes in these assumptions are independent and may move in either similar or opposite directions.

The significant unobservable inputs used in the fair value measurement of our derivative assets, derivative liabilities and VIE debt relate primarily to projected losses. In addition, when determining the fair value of our liabilities, we are required to incorporate into the fair value of those liabilities an adjustment that reflects our own non-performance risk, if applicable, as discussed below.

For our corporate CDOs, we estimate the correlation of the default probability between the corporate entities and Radian—the higher the correlation percentage, the higher the probability that both the corporate entities and Radian will default together. In addition, a widening of the average credit spread increases the expected loss for our transactions, and therefore, increases the related liability.

For our CDO of CMBS transactions, we use the CMBX index that most directly correlates to our transaction with respect to vintage and credit rating and then we estimate losses by applying a correlation factor. Because we have more exposure to senior tranches, an increase in this factor generally increases the expected loss for our transactions and therefore, increases our related liability.

For our TruPs CDOs, the performance of each underlying reference obligation is measured by a standard and distressed pricing, which indicates the expected principal recovery. An increase in the standard and stressed principal recovery decreases the loss severity of the transaction, and therefore, in isolation, decreases the related liability. For those transactions where we may be required to pay a “liquidity claim,” we also assign these transactions a probability that we will be required to pay such claim, which generally would increase our related liability. See “Item 7. Management’s Discussion and Analysis of Financial Condition and Results of Operations—Liquidity and Capital Resources—Financial Guaranty” for additional information regarding the potential payment of a liquidity claim.

For our TruPs-related VIE liabilities, the fair value is estimated using similar inputs as in the estimated fair value of our TruPs CDOs, except there is no non-performance risk adjustment, as the derivative liability is limited to the segregated assets already held by the VIE.

For our other non-corporate CDOs, we utilize the internal credit rating, average remaining life, and current par outstanding for each transaction to project both expected losses and an internally developed risk-based capital amount. An increase in the average remaining life typically increases the expected loss of the transactions, and therefore, increases our related liability. An upgrade (downgrade) in the internal credit rating typically decreases (increases) the expected loss of the transactions, and therefore, decreases (increases) our related liability.

For all fair value measurements where we project our non-performance risk, including VIE debt, we utilize a market observed credit spread for Radian, which we believe is the best available indicator of the market’s perception of our non-performance risk. In isolation, a widening (tightening) of this credit spread typically decreases (increases) our related liability. The assumption used to project our own non-performance risk is independent from the other unobservable inputs used in our fair value measurements. The net impact on our reported assets and liabilities from increases or decreases in our own credit spread and from increases or decreases in other unobservable inputs depends upon the magnitude and direction of the changes in each input; such changes may result in offsetting effects to our recorded fair value measurements or they may result in directionally similar impacts, which may be material.

Corporate CDOs

The fair value of each of our corporate CDO transactions is estimated based on the difference between: (1) the present value of the expected future contractual premiums we charge; and (2) the fair premium amount that we estimate that another financial guarantor would require to assume the rights and obligations under our contracts. The fair value estimates reflect the fair value of the asset or liability, which is consistent with the “in-exchange” approach, in which fair value is determined based on the price that would be received or paid in a current transaction as defined by the accounting standard regarding fair value measurements. These credit derivatives are categorized in Level III of the fair value hierarchy.

Present Value of Expected Future Contractual Premiums—Our contractual premiums are subject to change primarily for two reasons: (1) all of our contracts provide our counterparties with the right to terminate upon our default; and (2) 88% of the aggregate net par outstanding of our corporate CDO transactions (as of December 31, 2013) provide our counterparties with an additional right to terminate these transactions that is currently exercisable at any time. In determining the expected future premiums of these transactions, we adjust the contractual premiums for such transactions to reflect the estimated fair value of those premiums based on our estimate of the probability of our counterparties exercising this termination right and the impact it would have on the remaining expected lifetime premium. We also cap the total estimated fair value of the contracts subject to termination such that none of these contracts are in a derivative asset position. As of December 31, 2013, 8% of the aggregate net par outstanding of our corporate CDO transactions was capped in this manner. The discount rate we use to determine the present value of expected future premiums is our CDS spread plus a risk-free rate. This discount rate reflects the risk that we may not collect future premiums due to our inability to satisfy our contractual obligations, which provides our counterparties the right to terminate the contracts.

Determining the Fair Premium Amount—For each corporate CDO transaction, we perform three principal steps in determining the fair premium amount:


•	first, we define a tranche on the CDX index (defined below) that equates to the risk profile of our specific transaction (we refer to this tranche as an “equivalent-risk tranche”);


•	second, we determine the fair premium amount on the equivalent-risk tranche for those market participants engaged in trading on the CDX index (we refer to each of these participants as a “typical market participant”); and


•	third, we adjust the fair premium amount for a typical market participant to account for the difference between the non-performance or default risk of a typical market participant and the non-performance or default risk of a financial guarantor of similar credit quality to us (in each case, we refer to the risk of non-performance as “non-performance risk”).

Defining the Equivalent-Risk Tranche—Direct observations of fair premium amounts for our transactions are not available because these transactions cannot be traded or transferred pursuant to their terms and there is currently no active market for these transactions. However, CDS on tranches of a standardized index (the “CDX index”) are widely traded and observable and provide relevant market data for determining the fair premium amount of our transactions, as described more fully below.

The CDX index is an index based on a synthetic corporate CDO that comprises a list of corporate obligors and is segmented into multiple tranches of synthetic senior unsecured debt of these obligors ranging from the equity tranche (i.e., the most credit risk or first-loss position) to the most senior tranche (i.e., the least credit risk). We refer to each of these tranches as a “standard CDX tranche.” A tranche is defined by an attachment point and detachment point, representing the range of portfolio losses for which the protection seller would be required to make a payment.

Our corporate CDO transactions possess similar structural features to the standard CDX tranches, but often differ with respect to the referenced corporate entities, the term, the attachment points and the detachment points. Therefore, in order to determine the equivalent-risk tranche for each of our corporate CDO transactions, we determine the attachment and detachment points on the CDX index that have comparable estimated probabilities of loss as the attachment and detachment points in our transactions. We begin by performing a simulation analysis of referenced entity defaults in our transactions to determine the probability of portfolio losses exceeding our attachment and detachment points. The referenced entity defaults are primarily determined based on the following inputs: the market observed CDS credit spreads of the referenced corporate entities, the correlations between each of the referenced corporate entities and the term of the transaction.

For each referenced corporate entity in our corporate CDO transactions, the CDS spreads associated with the term of our transactions (“credit curve”) define the estimated expected loss for each entity (as applied in a market standard approach known as “risk neutral” modeling). The credit curves on individual referenced entities are generally observable. The expected cumulative loss for the portfolio of referenced entities associated with each of our transactions is the sum of the expected losses of these individual referenced entities. With respect to the correlation of losses across the underlying reference entities, two obligors belonging to the same industry or located in the same geographical region are assumed to have a higher probability of defaulting together (i.e., they are more correlated). An increase in the correlations between the referenced entities generally causes a higher expected loss for the portfolio associated with our transactions. The estimated correlation factors that we use are derived internally based on observable third-party inputs from historical data.

The impact of our correlation assumptions currently does not have a material effect on our fair premium estimates in light of the significant impact of our non-performance risk adjustment as described below.

Once we have established the probability of portfolio losses exceeding the attachment and detachment points in our transactions, we then use the same simulation method to locate the attachment and detachment points on the CDX index with comparable probabilities. These equivalent attachment and detachment points define the equivalent-risk tranche on the CDX index that we use to determine fair premium amounts.

Determining the Typical Fair Premium Amount—The equivalent-risk tranches for our corporate CDO transactions often are not identical to any standard CDX tranches. As a result, fair premium amounts generally are not directly observable from the CDX index for the equivalent-risk tranche and must be separately determined. We make this determination through an interpolation in which we use the observed premium rates on the standard CDX tranches that most closely match our equivalent-risk tranche to derive the typical fair premium amount for the equivalent-risk tranche.

Non-Performance Risk Adjustment on Corporate CDOs—The typical fair premium amount estimated for the equivalent-risk tranche represents the fair premium amount for a typical market participant—not Radian. Accordingly, the final step in our fair value estimation is to convert this typical fair premium amount into a fair premium amount for a financial guarantor of similar credit quality to us. A typical market participant is contractually bound by a requirement that collateral be posted regularly to minimize the impact of that participant’s default or non-performance. This collateral posting feature makes these transactions less risky to the protection buyer, and therefore, priced differently. None of our contracts require us to post collateral with our counterparties, which exposes our counterparties fully to our non-performance risk. We make an adjustment to the typical fair premium amount to account for both this contractual difference, as well as for the market’s perception of our default probability, which is observable through our CDS spread.

The amount of the non-performance risk adjustment is computed based, in part, on the expected claim payment by Radian. To estimate this expected payment, we first determine the expected claim payment of a typical market participant by using a risk-neutral modeling approach. A significant underlying assumption of the risk-neutral model approach that we use is that the typical fair premium amount is equal to the present value of expected claim payments from a typical market participant. Expected claim payments on a transaction are based on the expected loss on that transaction (also determined using the risk-neutral modeling approach). Radian’s expected claim payment is calculated based on the correlation between the default probability of the transaction and our default probability. The default probability of Radian is determined from the observed Radian Group CDS spread and the default probability of the transaction is determined as described above under “—Defining the Equivalent-Risk Tranche.” The present value of Radian’s expected claim payments is discounted using a risk-free interest rate, as the expected claim payments have already been risk-adjusted.

Our non-performance risk adjustment currently results in a material reduction of our typical fair premium amounts, which in turn has a positive impact on the fair value of these derivatives.

Non-Corporate CDOs and Other Derivative Transactions

Our non-corporate CDO transactions include our guaranty of TruPs CDOs, CDOs of CMBS and CDOs backed by other asset classes such as: (i) municipal securities; (ii) synthetic financial guarantees of ABS; and (iii) project finance transactions. The fair value of our non-corporate CDOs and other derivative transactions is calculated as the difference between the present value of the expected future contractual premiums and our estimate of the fair premium amount for these transactions. The present value of expected future contractual premiums is determined based on the methodology described above for corporate CDOs. As of December 31, 2013, 86% of the aggregate net par outstanding of our non-corporate CDO contracts provide our counterparties with the right to terminate these transactions that is exercisable at any time. We also cap the total estimated fair value of the contracts subject to termination such that none of these contracts are in a derivative asset position. As of December 31, 2013, 26% of the aggregate net par outstanding of our non-corporate CDO transactions was capped in this manner. In all other instances, we utilize internal models to estimate the fair premium amount as described below. These credit derivatives are categorized in Level III of the fair value hierarchy.

TruPs CDOs and TruPs-Related VIE Liabilities—Our TruPs transactions are CDS on CDOs where the collateral consists primarily of deeply subordinated securities issued by banks, insurance companies, real estate investment trusts and other financial institutions whose individual spreads are not observable. In each case, we provide credit protection on a specific tranche of each CDO. To determine fair value for these transactions, we use a discounted cash flow valuation approach that captures the credit characteristics of each transaction. We estimate projected claims based on our internal credit analysis, which is based on the current performance of each underlying reference obligation. The present value of the expected cash flows of the TruPs transaction is then determined using a discount rate derived from the observed market pricing for a TruPs transaction with similar characteristics. The present value of the insured cash flows is determined using a discount rate that is equal to our CDS rate plus a risk-free rate.

With respect to four insured TruPs transactions, our counterparties may require that we pay a liquidity claim for the outstanding par amount on the underlying TruPs bonds. For these transactions, an additional fair value adjustment is made. To calculate this adjustment, a probability that we will be required to pay a liquidity claim is assigned based on our internal cash flow projections. A discounted cash flow valuation is also performed for this scenario where we are required to pay a liquidity claim. The fair value is set equal to the probability weighted average of the valuations from two scenarios: one in which our counterparty pays a liquidity claim and one in which the claim is not paid.

CDOs of CMBS—The fair premium amounts for our CDO of CMBS transactions for a typical market participant are derived first by observing the spreads of the CMBX indices that match the underlying reference obligations of our transactions. A mezzanine tranche, which represents our insured tranche, is then priced through a standard CDO model. The CMBX indices represent standardized lists of CMBS reference obligations. A different CMBX index exists for different types of underlying referenced obligations based on vintages and credit rating. For each of our CDO of CMBS transactions, we use the CMBX index that most directly correlates to our transaction with respect to vintage and credit rating. Because the observable CMBS indices do not have a similar mezzanine tranche, we use an internal CDO pricing model in order to adjust fair value for this structural feature. A standard CDO pricing model was calibrated to establish the market pricing at inception. This CDO pricing model is then applied to the current valuation period to derive the fair premium for the mezzanine tranche. The typical fair premium amount represents the estimated fair value of the expected future fair premiums determined by using a discount rate equal to the CDS spread of a typical market participant plus a risk-free rate.

All Other Non-Corporate CDOs and Other Derivative Transactions—The fair value estimates for certain derivative instruments are derived using observable market inputs that are indicative of exit price, when available, and are categorized in Level II of the fair value hierarchy. For all of our other non-corporate CDOs and other derivative transactions, observed prices and market indices are not available. As a result, we utilize an internal model that estimates fair premium. The fair premium amount is calculated such that the expected profit (fair premium amount net of expected losses and other expenses) is proportional to an internally-developed risk-based capital amount. Expected losses and our internally developed risk-based capital amounts are projected by our model using the internal credit rating, term and current par outstanding for each transaction.

For each of the non-corporate CDOs and other derivative transactions discussed above, with the exception of our TruPs transactions that are valued using a discounted cash flow analysis, we make an adjustment to the fair premium amounts, as described above under “—Non-Performance Risk Adjustment on Corporate CDOs,” to incorporate our own non-performance risk. The non-performance risk adjustment associated with our TruPs transactions is incorporated in the fair value as described above; therefore, no separate adjustment is required. These credit derivatives are categorized in Level III of the fair value hierarchy.

Assumed Financial Guaranty Credit Derivatives

In making our determination of fair value for these credit derivatives, we use information provided to us by our counterparties to these reinsurance transactions, which are the primary insurers (the “primaries”) of the underlying credits, including the primaries’ fair valuations for these credits. The information obtained from our counterparties is not received with sufficient time for us to properly record the mark-to-market liability as of the balance sheet date. Therefore, the amount recorded as of December 31, 2013 is based on the most recent available financial information, which is reported on a quarterly lag. The lag in reporting is consistent from period to period. The fair value is based on credit spreads obtained by the primaries from market data sources published by third parties (e.g., dealer spread tables for collateral similar to assets within the transactions being valued), as well as collateral-specific spreads provided by trustees or obtained from market sources if such data is available. If observable market spreads are not available or reliable for the underlying reference obligations, then the primaries’ valuations are predominantly based on market indices that most closely resemble the underlying reference obligations, considering asset class, credit quality rating and maturity of the underlying reference obligations. In addition, these valuations incorporate an adjustment for non-performance risk. The primaries’ models used to estimate the fair value of these instruments include a number of factors, including credit spreads, changes in interest rates and the credit ratings of referenced entities. In establishing our fair value for these transactions, we assess the reasonableness of the primaries’
valuations by: (1) reviewing the primaries’ publicly available information regarding their mark-to-market processes, including methodology and key assumptions; and (2) analyzing and discussing the changes in fair value with the primaries where the changes appear unusual or do not appear materially consistent with credit loss related information when provided by the primaries for these transactions. These credit derivatives are categorized in Level III of the fair value hierarchy.

Other Financial Guaranty VIE Consolidated Assets/Liabilities

We are the primary beneficiary for two other VIEs for which we have provided financial guarantees. These VIEs primarily consist of manufactured housing loans and VIE debt to note holders in the trust. The fair value of the VIE debt related to these other financial guaranty VIEs is estimated based on prices of comparable securities and spreads observed in the market. The overall net fair value for these transactions is determined using a discounted cash flow analysis. We do not currently estimate any projected claims based on our internal credit analysis, which is based on the current performance of the underlying collateral and the remaining subordination available to support the transaction. The present value of the insured cash flows is determined by using a discount rate that is equal to our CDS rate plus a risk-free rate. We utilize this model to determine the fair value of our exposure to these VIEs and to derive the fair value of the assets in these VIEs, which are reported within other assets on our consolidated balance sheets.

Fixed-Maturities Held to Maturity—The fair values of fixed-maturity securities are obtained from independent pricing services that use observed market transactions, including broker-dealer quotes and actual trade activity as a basis for valuation.

Other Invested Assets—The fair value of these assets, primarily invested in limited partnerships, is estimated based on information within the financial statements provided by the limited partnerships. These interests are accounted for and carried as cost-method investments.

Long-Term Debt—The carrying amount of long-term debt is net of the equity component of our convertible notes, which is accounted for under the accounting standard for convertible debt instruments that may be settled in cash upon conversion (including partial cash settlement). The fair value is estimated based on the quoted market prices for the same or similar issues. See Note 11 for further information.

Non-Derivative Financial Guaranty Liabilities—We estimate the fair value of these non-derivative financial guarantees in a hypothetical market where market participants include other monoline mortgage and financial guaranty insurers with similar credit quality to us, assuming that the net liability related to these insurance contracts could be transferred to these other mortgage and financial guaranty insurance and reinsurance companies.

This fair value estimate of non-derivative financial guarantees includes direct and assumed contracts written and is based on the difference between the present value of: (1) the expected future contractual premiums; and (2) the fair premium amount to provide the same credit protection assuming a transfer of our obligation to a guarantor of similar credit quality as Radian as of the measurement date.

The key variables considered in estimating fair value include par amounts outstanding (including future periods for the estimation of future installment premiums), expected term, unearned premiums, expected losses and our CDS spread. Estimates of future installment premiums received are based on contractual premium rates.

With respect to the fair premium amount, the accounting standard regarding fair value measurements requires that the non-performance risk of a financial liability be included in the estimation of fair value. Accordingly, the fair premium amount for financial guaranty insurance contracts includes consideration of our credit quality as represented by our CDS spread.

Insurance Premiums-Revenue Recognition

Mortgage insurance premiums written on an annual and multi-year basis are initially recorded as unearned premiums and earned over time. Annual premiums are amortized on a monthly, straight-line basis. Multi-year premiums are amortized over the terms of the contracts in relation to the anticipated claim payment pattern based on historical industry experience. Premiums written on a monthly basis are earned over the period that coverage is provided. When we rescind insurance coverage on a loan, we refund all premiums received in connection with such coverage. Premium revenue is recognized net of our accrual for estimated rescission refunds. With respect to our reinsurance transactions, ceded premiums written are initially set up as prepaid reinsurance and are amortized in a manner consistent with the recognition of income on direct premiums. Premiums on certain structured transactions in our mortgage insurance business are recognized over the period that coverage is provided.

In our financial guaranty business, insurance premiums are earned in proportion to the level of amortization of insured principal over the contract period or over the period that coverage is provided. Unearned premiums represent that portion of premiums that will be earned over the remainder of the contract period. We record the initial unearned premium liability on installment policies equal to the present value of the premiums due or expected to be collected over either the period of the policy or, for certain sectors with a homogeneous pool of assets underlying the insured financial obligation, the expected period of risk. In determining the present value of premiums due, we use a discount rate that reflects a risk-free rate. Premiums paid in full at inception are recorded as unearned premiums. In addition, we recognize the remaining unearned premium revenue when bonds issued are redeemed or otherwise retired (“refundings”) that results in the extinguishment of the financial guaranty policies insuring such bonds. A refunding that is effected through the deposit of cash or permitted securities into an irrevocable trust for repayment, when permitted under the applicable bond indenture (a “legal defeasance”), does not qualify for immediate revenue recognition since the defeased obligation legally remains outstanding and covered by our insurance. See Note 10 for further information. Assumed premiums in reinsurance transactions are based on information reported by ceding companies. Premiums ceded through reinsurance agreements reduce premiums earned in a manner that is consistent with the recognition of the gross premiums as described above. When insured obligations are refunded or called, the remaining premiums are generally earned at that time.

Credit enhancement fees earned on derivative contracts are included in the change in fair value of derivative instruments.

Deferred Policy Acquisition Costs

Incremental, direct costs associated with the acquisition of mortgage insurance business, consisting of compensation and other policy issuance and underwriting expenses, are initially deferred and reported as deferred policy acquisition costs. Amortization of these costs for each underwriting year book of business is charged against revenue in proportion to estimated gross profits over the estimated life of the policies. This includes accruing interest on the unamortized balance of deferred policy acquisition costs. Ceding commissions received under our reinsurance agreements related to these costs are also deferred and accounted for using similar assumptions. Estimates of expected gross profit, including persistency and loss development assumptions for each underwriting year used as a basis for amortization, are evaluated quarterly and the total amortization recorded to date is adjusted by a charge or credit to our consolidated statements of operations if actual experience or other evidence suggests that previous estimates should be revised. Considerable judgment is used in evaluating these estimates and the assumptions on which they are based. The use of different assumptions may have a significant effect on the amortization of deferred policy acquisition costs.

Effective January 1, 2012, we adopted the FASB update to the accounting standard regarding accounting for costs associated with acquiring or renewing insurance contracts on a prospective basis. This update redefines acquisition costs as incremental costs that are related directly to the successful acquisition of new or renewal insurance contracts. Previously, acquisition costs were defined as costs that vary with and are primarily related to the acquisition of insurance contracts. The effect of this revised definition of acquisition costs has resulted in the recognition of additional expenses in our mortgage insurance business when incurred, rather than being deferred to subsequent periods. There was no change to the amortization requirements due to this update. The implementation of this new guidance significantly reduced the amount of our deferral of policy acquisition costs associated with acquiring mortgage insurance contracts. However, the lower amount of acquisition costs deferred will be offset by reduced amortization expense in subsequent periods. While the timing of the recognition of certain costs in our results of operations has changed as a result of the adoption of this update, there is no effect on the total acquisition costs recognized over time or on our cash flows. Amounts deferred as acquisition costs for 2013 and 2012 reflect a reduction for ceding commissions written on risk ceded under the Reinsurance Transactions (as defined in Note 8). We amortized $28.5 million, $34.1 million and $36.1 million of deferred policy acquisition costs in our mortgage insurance business in 2013, 2012 and 2011, respectively.

The adoption of the FASB update to the accounting standard for deferred acquisition costs did not impact the deferred acquisition costs for the financial guaranty business, because we were not writing new financial guaranty business subsequent to our prospective implementation of the new guidance. Deferred policy acquisition costs in the financial guaranty business are comprised of those expenses that vary with, and are principally related to, the production of insurance premiums, including: commissions paid on reinsurance assumed, salaries and related costs of underwriting and marketing personnel, rating agency fees, premium taxes and certain other underwriting expenses, offset by commission income on premiums ceded to reinsurers. Acquisition costs are deferred and amortized over the period in which the related premiums are earned for each underwriting year. The estimation of installment-based premiums requires considerable judgment, and different assumptions could produce different results. We amortized $13.2 million, $27.7 million, and $16.7 million of deferred policy acquisition costs in our financial guaranty business during 2013, 2012 and 2011, respectively.

Income Taxes

We provide for income taxes in accordance with the provisions of the accounting standard regarding accounting for income taxes. As required under this standard, our deferred tax assets and liabilities are recognized under the balance sheet method, which recognizes the future tax effect of temporary differences between the amounts recorded in our consolidated financial statements and the tax bases of these amounts. Deferred tax assets and liabilities are measured using the enacted tax rates expected to apply to taxable income in the periods in which the deferred tax asset or liability is expected to be realized or settled.

We are required to establish a valuation allowance against our deferred tax asset (“DTA”) when it is more likely than not that all or some portion of our DTA will not be realized. At each balance sheet date, we assess our need for a valuation allowance. Our assessment is based on all available evidence, both positive and negative. This requires management to exercise judgment and make assumptions regarding whether such DTA will be realized in future periods. Future realization of our DTA will ultimately depend on the existence of sufficient taxable income of the appropriate character (ordinary income or capital gains) within the applicable carryback and carryforward periods provided under the tax law. In making this assessment, the primary negative evidence that we considered are our cumulative losses in recent years and the continued uncertainty around our future results. We also considered positive evidence when assessing the need for a valuation allowance, such as future reversals of existing taxable temporary differences, future projections of taxable income, taxable income within the applicable carryback periods and potential tax planning strategies. In assessing our need for a valuation allowance, the weight assigned to the effect of both negative and positive evidence is commensurate with the extent to which such evidence can be objectively verified.

Given the impact on our pretax results of net gains or losses from our derivative transactions and investment portfolio, and the continued uncertainty around our ability to rely on short-term financial projections, which directly affects our ability to estimate an effective tax rate for the full year, we booked our income tax expense (benefit) for interim periods based on actual results of operations.

Foreign Currency Revaluation/Translation

Assets and liabilities denominated in foreign currencies are revalued or translated at year-end exchange rates. Operating results are translated at average rates of exchange prevailing during the year. Unrealized gains and losses, net of deferred taxes, resulting from translation are included in accumulated other comprehensive income (loss) in stockholders’ equity. Realized gains and losses resulting from transactions in foreign currency are recorded in our statements of operations.

Cash and Restricted Cash

Included in our restricted cash balances as of December 31, 2013 were: (1) funds for a mortgage insurance reserve policy held in escrow for any future duties, rights and liabilities; (2) funds held as collateral under our insurance trust agreements related to health care benefits; and (3) funds held in trust for the benefit of certain policyholders.

Within our consolidated statements of cash flows, we classify cash receipts and cash payments related to items measured at fair value according to their nature and purpose. While our securities trading activity was significant in 2012 and 2011, this activity was primarily driven by strategic repositioning of the portfolio in order to: (1) shorten duration for liquidity purposes; and (2) increase our allocation to taxable bonds to maximize our after-tax yields. Because this activity relates to overall strategic initiatives and is not trading related, it is reflected as cash flows from investing activities.

Investments

We group assets in our investment portfolio into one of three main categories: held to maturity, available for sale or trading securities. Fixed-maturity securities for which we have the positive intent and ability to hold to maturity are classified as held to maturity and are reported at amortized cost. Investments in securities not classified as held to maturity or trading securities are classified as available for sale and are reported at fair value, with unrealized gains and losses (net of tax) reported as a separate component of stockholders’ equity as accumulated other comprehensive income (loss). Investments classified as trading securities are reported at fair value, with unrealized gains and losses reported as a separate component of income. Short-term investments consist of money market instruments, certificates of deposit (“CDs”) and highly liquid, interest-bearing instruments with an original maturity of three months or less at the time of purchase. Amortization of premium and accretion of discount are calculated principally using the interest method over the term of the investment. Realized gains and losses on investments are recognized using the specific identification method. See Note 5 for further discussion on the fair value of investments.

For certain hybrid financial instruments that would be required to be separated into a host contract and a derivative instrument, the accounting standard regarding derivatives and hedging permits an entity to irrevocably elect to initially and subsequently measure that hybrid financial instrument in its entirety at fair value (with changes in fair value recognized in earnings). We elected to record our convertible securities meeting these criteria at fair value with changes in the fair value recorded as net gains or losses on investments. All hybrid financial instruments are classified as trading securities.

We record an OTTI adjustment on a security if we intend to sell the impaired security, if it is more likely than not that we will be required to sell the impaired security prior to recovery of its amortized cost basis, or if the present value of cash flows we expect to collect is less than the amortized cost basis of the security. If a sale is likely, the security is classified as other-than-temporarily impaired and the full amount of the impairment is recognized as a loss in the statement of operations. Otherwise, losses on securities that are other-than-temporarily impaired are separated into: (i) the portion of loss that represents the credit loss; and (ii) the portion that is due to other factors. The credit loss portion is recognized as a loss in the statement of operations, while the loss due to other factors is recognized in accumulated other comprehensive income (loss), net of taxes. A credit loss is determined to exist if the present value of discounted cash flows expected to be collected from the security is less than the cost basis of the security. The present value of discounted cash flows is determined using the original yield of the security. In evaluating whether a decline in value is other-than-temporary, we consider several factors in addition to the above, including, but not limited to, the following:


•	the extent and the duration of the decline in value;


•	the reasons for the decline in value (e.g., credit event, interest related or market fluctuations); and


•	the financial position, access to capital and near term prospects of the issuer, including the current and future impact of any specific events.

Accounts and Notes Receivable

Accounts and notes receivable primarily consist of accrued premiums receivable due from our mortgage insurance and financial guaranty customers. Accounts and notes receivable are carried at their estimated collectible amounts, net of any allowance for doubtful accounts, and are periodically evaluated for collectability based on past payment history and current economic conditions.

Company-Owned Life Insurance

We are the beneficiary of insurance policies on the lives of certain of our current and past officers and employees. We have recognized the amount that could be realized upon surrender of the insurance policies in other assets in our consolidated balance sheets. At December 31, 2013 and 2012, the cash surrender value of company-owned life insurance totaled $78.4 million and $76.5 million, respectively.

Property and Equipment

Property and equipment is carried at cost, net of depreciation. For financial statement reporting purposes, computer hardware and software is depreciated over three years and furniture, fixtures and office equipment is depreciated over seven years. Leasehold improvements are depreciated over the lesser of the life of the asset improved or the remaining term of the lease. For income tax purposes, we use accelerated depreciation methods.

Accounting for Stock-Based Compensation

The stock-based compensation cost related to share-based liability awards is based on the fair value as of the measurement date. The compensation cost for equity instruments is measured based on the grant-date fair value at the date of issuance. Compensation cost is recognized over the periods that an employee provides service in exchange for the award. See Note 15 for further information.

We use the Monte Carlo valuation model to determine the fair value of all cash-settled awards where stock price is a factor in determining the vesting, as well as for cash or stock settled performance awards where there exists a similar stock price based market condition (we refer to these awards as “Market Condition Awards”). The Monte Carlo valuation model incorporates multiple input variables, including expected life, volatility, risk-free rate of return and dividend yield for each award to estimate the probability that a vesting condition will be achieved. In determining these assumptions for the Monte Carlo valuations, we consider historic and observable market data.

Depending on certain characteristics of the awards granted under the various compensation programs noted above, they are accounted for as either liabilities or equity instruments.

In 2013 and 2012, we used the Monte Carlo valuation model in determining the grant date fair value of performance-based stock options issued to executives and non-executives. Prior to 2012, we used the Black-Scholes valuation model in determining the grant date fair value of all stock options issued to executives, non-executives and non-employee directors.

We elected to apply the short-cut method in accounting for the windfall tax benefits under the accounting standard regarding share-based payment.