Fair Value Measurement

The Company carries a significant portion of its assets and liabilities at fair value. Fair value is defined as the price 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 (i.e., exit price). The price represents the price available in the principal market for the asset or liability. If there is no principal market, then the price is based on a hypothetical market that maximizes the value received for an asset or minimizes the amount paid for a liability (i.e., the most advantageous market).

Fair value is based on quoted market prices, where available. If listed prices or quotes are not available, fair value is based on either internally developed models that primarily use, as inputs, market-based or independently sourced market parameters, including but not limited to yield curves, interest rates and debt prices or with the assistance of an independent third-party using a discounted cash flow approach and the third party’s proprietary pricing models. In addition to market information, models also incorporate transaction details, such as maturity of the instrument and contractual features designed to reduce the Company’s credit exposure, such as collateral rights as applicable.

Valuation adjustments may be made to ensure that financial instruments are recorded at fair value. These adjustments include amounts to reflect counterparty credit quality, the Company’s creditworthiness and constraints on liquidity. As markets and products develop and the pricing for certain products becomes more or less transparent, the Company may refine its methodologies and assumptions. During 2018, no changes were made to the Company’s valuation models that had or are expected to have, a material impact on the Company’s consolidated balance sheets or statements of operations and comprehensive income.

The Company’s methods for calculating fair value produce a fair value that may not be indicative of net realizable value or reflective of future fair values. The use of different methodologies or assumptions to determine fair value of certain financial instruments could result in a different estimate of fair value at the reporting date.

The categorization within the fair value hierarchy is determined based on whether the inputs to valuation techniques used to measure fair value are observable or unobservable. Observable inputs reflect market data obtained from independent sources, while unobservable inputs reflect Company estimates of market assumptions. The fair value hierarchy prioritizes model inputs into three broad levels as follows, with Level 1 being the highest and Level 3 the lowest. An asset's or liability’s categorization is based on the lowest level of significant input to its valuation.

Level 1—Quoted prices for identical instruments in active markets. The Company generally defines an active market as a market in which trading occurs at significant volumes. Active markets generally are more liquid and have a lower bid-ask spread than an inactive market.

Level 2—Quoted prices for similar instruments in active markets; quoted prices for identical or similar instruments in markets that are not active; and observable inputs other than quoted prices, such as interest rates or yield curves and other inputs derived from or corroborated by observable market inputs.

Level 3—Model derived valuations in which one or more significant inputs or significant value drivers are unobservable. Financial instruments are considered Level 3 when their values are determined using pricing models, discounted cash flow methodologies or similar techniques and at least one significant model assumption or input is unobservable. Level 3 financial instruments also include those for which the determination of fair value requires significant management judgment or estimation.

During the periods presented, there were no transfers into or from Level 3 except for one transfer from Level 2 into Level 3 during 2017 because starting in the second quarter of 2017 the price of the security includes a significant unobservable assumption.

Measured and Carried at Fair Value

Fixed-Maturity Securities and Short-Term Investments

The fair value of bonds in the investment portfolio is generally based on prices received from third party pricing services or alternative pricing sources with reasonable levels of price transparency. The pricing services prepare estimates of fair value measurements using their pricing models, which take into account: benchmark yields, reported trades, broker/dealer quotes, issuer spreads, two-sided markets, benchmark securities, bids, offers, reference data, industry and economic events and sector groupings. Additional valuation factors that can be taken into account are nominal spreads and liquidity adjustments. The pricing services evaluate each asset class based on relevant market and credit information, perceived market movements, and sector news.

Benchmark yields have in many cases taken priority over reported trades for securities that trade less frequently or those that are distressed trades, and therefore may not be indicative of the market. The extent of the use of each input is dependent on the asset class and the market conditions. The valuation of fixed-maturity investments is more subjective when markets are less liquid due to the lack of market based inputs.

Short-term investments that are traded in active markets are classified within Level 1 in the fair value hierarchy and their value is based on quoted market prices. Securities such as discount notes are classified within Level 2 because these securities are typically not actively traded due to their approaching maturity and, as such, their cost approximates fair value.

As of December 31, 2018, the Company used models to price 123 securities, primarily securities that were purchased or obtained for loss mitigation or other risk management purposes, with a fair value of $1,411 million. Most Level 3 securities were priced with the assistance of an independent third-party. The pricing is based on a discounted cash flow approach using the third-party’s proprietary pricing models. The models use inputs such as projected prepayment speeds; severity assumptions; recovery lag assumptions; estimated default rates (determined on the basis of an analysis of collateral attributes, historical collateral performance, borrower profiles and other features relevant to the evaluation of collateral credit quality); home price appreciation/depreciation rates based on macroeconomic forecasts and recent trading activity. The yield used to discount the projected cash flows is determined by reviewing various attributes of the security including collateral type, weighted average life, sensitivity to losses, vintage, and convexity, in conjunction with market data on comparable securities. Significant changes to any of these inputs could have materially changed the expected timing of cash flows within these securities which is a significant factor in determining the fair value of the securities.

Other Invested Assets

As of December 31, 2018 and December 31, 2017, other invested assets included investments carried and measured at fair value on a recurring basis of $3 million and $48 million, respectively. December 31, 2017 included primarily preferred stock investments in the global property catastrophe risk market and in a fund that invested primarily in senior loans and bonds, which were sold in 2018. Fair values for the preferred stock investments were based on their respective net asset value (NAV) per share or equivalent. Included in the amounts above are other equity investments that were carried at their fair value of $2 million as of December 31, 2018 and December 31, 2017. These equity investments were classified as Level 3.

Other Assets

Committed Capital Securities

The fair value of committed capital securities (CCS), which is recorded in “other assets” on the consolidated balance sheets, represents the difference between the present value of remaining expected put option premium payments under AGC CCS and AGM’s Committed Preferred Trust Securities (the AGM CPS) agreements, and the estimated present value that the Company would hypothetically have to pay currently for a comparable security (see Note 16, Long Term Debt and Credit Facilities). The AGC CCS and AGM CPS are carried at fair value with changes in fair value recorded in other income in the consolidated statement of operations. The estimated current cost of the Company’s CCS is based on several factors, including AGM and AGC CDS spreads, London Interbank Offered Rate (LIBOR) curve projections, the Company's publicly traded debt and the term the securities are estimated to remain outstanding.

Supplemental Executive Retirement Plans

The Company classifies the fair value measurement of the assets of the Company's various supplemental executive retirement plans as either Level 1 or Level 2. The fair value of these assets is valued based on the observable published daily values of the underlying mutual fund included in the aforementioned plans (Level 1) or based upon the NAV of the funds if a published daily value is not available (Level 2). The NAV's are based on observable information.

Contracts Accounted for as Credit Derivatives

The Company’s credit derivatives primarily consist of insured CDS contracts, and also include interest rate swaps that fall under derivative accounting standards requiring fair value accounting through the statement of operations. The following is a description of the fair value methodology applied to the Company's insured CDS that are accounted for as credit derivatives. The Company did not enter into CDS with the intent to trade these contracts and the Company may not unilaterally terminate a CDS contract absent an event of default or termination event that entitles the Company to terminate such contracts; however, the Company has mutually agreed with various counterparties to terminate certain CDS transactions. In transactions where the counterparty does not have the right to terminate, such transactions are generally terminated for an amount that approximates the present value of future premiums or for a negotiated amount, rather than at fair value.

The terms of the Company’s CDS contracts differ from more standardized credit derivative contracts sold by companies outside the financial guaranty industry. The non-standard terms generally include the absence of collateral support agreements or immediate settlement provisions. In addition, the Company employs relatively high attachment points and does not exit derivatives it sells, except under specific circumstances such as mutual agreements with counterparties. Management considers the non-standard terms of its credit derivative contracts in determining the fair value of these contracts.

Due to the lack of quoted prices and other observable inputs for its instruments or for similar instruments, the Company determines the fair value of its credit derivative contracts primarily through internally developed, proprietary models that use both observable and unobservable market data inputs. There is no established market where financial guaranty insured credit derivatives are actively traded; therefore, management has determined that the exit market for the Company’s credit derivatives is a hypothetical one based on its entry market. Management has tracked the historical pricing of the Company’s transactions to establish historical price points in the hypothetical market that are used in the fair value calculation. These contracts are classified as Level 3 in the fair value hierarchy as there are multiple unobservable inputs deemed significant to the valuation model, most importantly the Company’s estimate of the value of the non-standard terms and conditions of its credit derivative contracts and how the Company’s own credit spread affects the pricing of its transactions.

The fair value of the Company’s credit derivative contracts represents the difference between the present value of remaining premiums the Company expects to receive or pay and the estimated present value of premiums that a financial guarantor of comparable credit-worthiness would hypothetically charge or pay at the reporting date for the same protection. The fair value of the Company’s credit derivatives depends on a number of factors, including notional amount of the contract, expected term, credit spreads, changes in interest rates, the credit ratings of referenced entities, the Company’s own credit risk and remaining contractual cash flows. The expected remaining contractual premium cash flows are the most readily observable inputs since they are based on the CDS contractual terms. Credit spreads capture the effect of recovery rates and performance of underlying assets of these contracts, among other factors. Consistent with previous years, market conditions at December 31, 2018 were such that market prices of the Company’s CDS contracts were not available.

Assumptions and Inputs

The various inputs and assumptions that are key to the establishment of the Company’s fair value for CDS contracts are as follows: the gross spread, the allocation of gross spread among the bank profit, net spread and hedge cost, and the weighted average life which is based on debt service schedules. The Company obtains gross spreads on its outstanding contracts from market data sources published by third parties (e.g., dealer spread tables for the collateral similar to assets within the Company’s transactions), as well as collateral-specific spreads provided by trustees or obtained from market sources. The bank profit represents the profit the originator, usually an investment bank, realizes for structuring and funding the transaction; the net spread represents the premiums paid to the Company for the Company’s credit protection provided; and the hedge cost represents the cost of CDS protection purchased by the originator to hedge its counterparty credit risk exposure to the Company.

With respect to CDS transactions for which there is an expected claim payment within the next twelve months, the allocation of gross spread reflects a higher allocation to the cost of credit rather than the bank profit component. In the current market, it is assumed that a bank would be willing to accept a lower profit on distressed transactions in order to remove these transactions from its financial statements.

Market sources determine credit spreads by reviewing new issuance pricing for specific asset classes and receiving price quotes from their trading desks for the specific asset in question. Management validates these quotes by cross-referencing quotes received from one market source against quotes received from another market source to ensure reasonableness. In addition, the Company compares the relative change in price quotes received from one quarter to another, with the relative change experienced by published market indices for a specific asset class. Collateral specific spreads obtained from third-party, independent market sources are un-published spread quotes from market participants or market traders who are not trustees. Management obtains this information as the result of direct communication with these sources as part of the valuation process. The following spread hierarchy is utilized in determining which source of gross spread to use.


•	Actual collateral specific credit spreads (if up-to-date and reliable market-based spreads are available).


•	Transactions priced or closed during a specific quarter within a specific asset class and specific rating. No transactions closed during the periods presented.


•	Credit spreads interpolated based upon market indices adjusted to reflect the non-standard terms of the Company's CDS contracts.


•	Credit spreads provided by the counterparty of the CDS.


•	Credit spreads extrapolated based upon transactions of similar asset classes, similar ratings, and similar time to maturity.

Information by Credit Spread Type (1)



	As of December 31, 2018		As of December 31, 2017
Based on actual collateral specific spreads	20	%	14	%
Based on market indices	33	%	48	%
Provided by the CDS counterparty	47	%	38	%
Total	100	%	100	%

____________________

(1) Based on par.

The rates used to discount future expected premium cash flows ranged from 2.47% to 2.89% at December 31, 2018 and 1.72% to 2.55% at December 31, 2017.

The premium the Company receives is referred to as the “net spread.” The Company’s pricing model takes into account not only how credit spreads on risks that it assumes affect pricing, but also how the Company’s own credit spread affects the pricing of its transactions. The Company’s own credit risk is factored into the determination of net spread based on the impact of changes in the quoted market price for credit protection bought on the Company, as reflected by quoted market prices on CDS referencing AGC or AGM. For credit spreads on the Company’s name the Company obtains the quoted price of CDS contracts traded on AGC and AGM from market data sources published by third parties. The cost to acquire CDS protection referencing AGC or AGM affects the amount of spread on CDS transactions that the Company retains and, hence, their fair value. As the cost to acquire CDS protection referencing AGC or AGM increases, the amount of premium the Company retains on a transaction generally decreases. Due to the low volume and total net par of CDS contracts remaining in AGM's portfolio, changes in AGM's credit spreads do not significantly affect the fair value of these CDS contracts.

In the Company’s valuation model, the premium the Company captures is not permitted to go below the minimum rate that the Company would currently charge to assume similar risks. This assumption can have the effect of mitigating the amount of unrealized gains that are recognized on certain CDS contracts. Given the current market conditions and the Company’s own credit spreads, approximately 17% and 16% based on fair value, of the Company's CDS contracts were fair valued using this minimum premium as of December 31, 2018 and December 31, 2017, respectively. The percentage of transactions that price using the minimum premiums fluctuates due to changes in AGC's credit spreads. In general when AGC's credit spreads narrow, the cost to hedge AGC's name declines and more transactions price above previously established floor levels. Meanwhile, when AGC's credit spreads widen, the cost to hedge AGC's name increases causing more transactions to price at previously established floor levels. The Company corroborates the assumptions in its fair value model, including the portion of exposure to AGC and AGM hedged by its counterparties, with independent third parties each reporting period. The current level of AGC’s and AGM’s own credit spread has resulted in the bank or transaction originator hedging a portion of its exposure to AGC and AGM. This reduces the amount of contractual cash flows AGC and AGM can capture as premium for selling its protection.

The amount of premium a financial guaranty insurance market participant can demand is inversely related to the cost of credit protection on the insurance company as measured by market credit spreads assuming all other assumptions remain constant. This is because the buyers of credit protection typically hedge a portion of their risk to the financial guarantor, due to the fact that the contractual terms of the Company's contracts typically do not require the posting of collateral by the guarantor. The extent of the hedge depends on the types of instruments insured and the current market conditions.

A credit derivative liability on protection sold is the result of contractual cash inflows on in-force transactions that are less than what a hypothetical financial guarantor could receive if it sold protection on the same risk as of the reporting date. If the Company were able to freely exchange these contracts (i.e., assuming its contracts did not contain proscriptions on transfer and there was a viable exchange market), it would realize a loss representing the difference between the lower contractual premiums to which it is entitled and the current market premiums for a similar contract. The Company determines the fair value of its CDS contracts by applying the difference between the current net spread and the contractual net spread for the remaining duration of each contract to the notional value of its CDS contracts and taking the present value of such amounts discounted at the corresponding LIBOR over the weighted average remaining life of the contract.

Strengths and Weaknesses of Model

The Company’s credit derivative valuation model, like any financial model, has certain strengths and weaknesses.

The primary strengths of the Company’s CDS modeling techniques are:


•	The model takes into account the transaction structure and the key drivers of market value.


•	The model maximizes the use of market-driven inputs whenever they are available.


•	The model is a consistent approach to valuing positions.

The primary weaknesses of the Company’s CDS modeling techniques are:


•	There is no exit market or any actual exit transactions; therefore, the Company’s exit market is a hypothetical one based on the Company’s entry market.


•	There is a very limited market in which to validate the reasonableness of the fair values developed by the Company’s model.


•	The markets for the inputs to the model are highly illiquid, which impacts their reliability.


•	Due to the non-standard terms under which the Company enters into derivative contracts, the fair value of its credit derivatives may not reflect the same prices observed in an actively traded market of credit derivatives that do not contain terms and conditions similar to those observed in the financial guaranty market.

Fair Value Option on FG VIEs’ Assets and Liabilities

The Company elected the fair value option for all the FG VIEs’ assets and liabilities and classifies them as Level 3 in the fair value hierarchy. The prices are generally determined with the assistance of an independent third-party, based on a discounted cash flow approach. The FG VIEs issued securities collateralized by first lien and second lien RMBS as well as loans and receivables.

The fair value of the Company’s FG VIEs’ assets is generally sensitive to changes in estimated prepayment speeds; estimated default rates (determined on the basis of an analysis of collateral attributes such as: historical collateral performance, borrower profiles and other features relevant to the evaluation of collateral credit quality); yields implied by market prices for similar securities; and house price depreciation/appreciation rates based on macroeconomic forecasts. Significant changes to some of these inputs could have materially changed the market value of the FG VIEs’ assets and the implied collateral losses within the transaction. In general, the fair value of the FG VIEs’ assets is most sensitive to changes in the projected collateral losses, where an increase in collateral losses typically could lead to a decrease in the fair value of FG VIEs’ assets, while a decrease in collateral losses typically leads to an increase in the fair value of FG VIEs’ assets. The third-party utilizes an internal model to determine an appropriate yield at which to discount the cash flows of the security, by factoring in collateral types, weighted-average lives, and other structural attributes specific to the security being priced. The expected yield is further calibrated by utilizing algorithms designed to aggregate market color, received by the independent third-party, on comparable bonds.

The models to price the FG VIEs’ liabilities used, where appropriate, the same inputs used in determining fair value of FG VIEs’ assets and, for those liabilities insured by the Company, the benefit from the Company's insurance policy guaranteeing the timely payment of principal and interest, taking into account the Company's own credit risk.

Significant changes to any of the inputs described above could have materially changed the timing of expected losses within the insured transaction which is a significant factor in determining the implied benefit from the Company’s insurance policy guaranteeing the timely payment of principal and interest for the tranches of debt issued by the FG VIEs that is insured by the Company. In general, extending the timing of expected loss payments by the Company into the future typically could lead to a decrease in the value of the Company’s insurance and a decrease in the fair value of the Company’s FG VIEs’ liabilities with recourse, while a shortening of the timing of expected loss payments by the Company typically could lead to an increase in the value of the Company’s insurance and an increase in the fair value of the Company’s FG VIEs’ liabilities with recourse.

Amounts recorded at fair value in the Company’s financial statements are presented in the tables below.

Fair Value Hierarchy of Financial Instruments Carried at Fair Value

As of December 31, 2018



			Fair Value Hierarchy
	Fair Value		Level 1		Level 2		Level 3
	(in millions)
Assets:
Investment portfolio, available-for-sale (1):
Fixed-maturity securities
Obligations of state and political subdivisions	$	4,911	$	—	$	4,812	$	99
U.S. government and agencies	175		—		175		—
Corporate securities	2,136		—		2,080		56
Mortgage-backed securities:
RMBS	982		—		673		309
Commercial mortgage-backed securities (CMBS)	539		—		539		—
Asset-backed securities	1,068		—		121		947
Non-U.S. government securities	278		—		278		—
Total fixed-maturity securities	10,089		—		8,678		1,411
Short-term investments	729		429		300		—
Other invested assets (2)	7		—		—		7
FG VIEs’ assets, at fair value (3)	569		—		—		569
Other assets (3) (4)	139		25		38		76
Total assets carried at fair value	$	11,533	$	454	$	9,016	$	2,063
Liabilities:
Credit derivative liabilities (3)	$	209	$	—	$	—	$	209
FG VIEs’ liabilities with recourse, at fair value (5)	517		—		—		517
FG VIEs’ liabilities without recourse, at fair value (3)	102		—		—		102
Total liabilities carried at fair value	$	828	$	—	$	—	$	828

Fair Value Hierarchy of Financial Instruments Carried at Fair Value

As of December 31, 2017



			Fair Value Hierarchy
	Fair Value		Level 1		Level 2		Level 3
	(in millions)
Assets:
Investment portfolio, available-for-sale (1):
Fixed-maturity securities
Obligations of state and political subdivisions	$	5,760	$	—	$	5,684	$	76
U.S. government and agencies	285		—		285		—
Corporate securities	2,018		—		1,951		67
Mortgage-backed securities:
RMBS	861		—		527		334
CMBS	549		—		549		—
Asset-backed securities	896		—		109		787
Non-U.S. government securities	305		—		305		—
Total fixed-maturity securities	10,674		—		9,410		1,264
Short-term investments	627		464		162		1
Other invested assets (2)	7		—		—		7
FG VIEs’ assets, at fair value (3)	700		—		—		700
Other assets (3) (4)	123		25		36		62
Total assets carried at fair value	$	12,131	$	489	$	9,608	$	2,034
Liabilities:
Credit derivative liabilities (3)	$	271	$	—	$	—	$	271
FG VIEs’ liabilities with recourse, at fair value (3)	627		—		—		627
FG VIEs’ liabilities without recourse, at fair value (3)	130		—		—		130
Total liabilities carried at fair value	$	1,028	$	—	$	—	$	1,028

____________________

(1) Change in fair value is included in OCI.


(2)	Excludes investments of $45 million as of December 31, 2017, measured using NAV per share with the change in fair value recorded in the consolidated statements of operations, which were sold in 2018. Includes Level 3 mortgage loans that are recorded at fair value on a non-recurring basis.

(3) Change in fair value is included in the consolidated statements of operations.

(4) Includes credit derivative assets.


(5)	Change in fair value attributable to ISCR is recorded in OCI with the remainder of the change in fair value recorded in the consolidated statements of operations.

Changes in Level 3 Fair Value Measurements

The tables below present a roll forward of the Company’s Level 3 financial instruments carried at fair value on a recurring basis during the years ended December 31, 2018 and 2017.

Fair Value Level 3 Rollforward

Recurring Basis

Year Ended December 31, 2018



	Fixed-Maturity Securities
	Obligations of State and Political Subdivisions					Corporate Securities					RMBS					Asset- Backed Securities					FG VIEs’ Assets at Fair Value					Other (7)					Credit Derivative Asset (Liability), net (5)					FG VIEs’ Liabilities with Recourse, at Fair Value					FG VIEs’ Liabilities without Recourse, at Fair Value
	(in millions)
Fair value as of December 31, 2017	$	76				$	67				$	334				$	787				$	700				$	64				$	(269	)			$	(627	)			$	(130	)
Total pretax realized and unrealized gains/(losses) recorded in: (1)
Net income (loss)	3			(2	)	(14		)	(2	)	21			(2	)	57			(2	)	2			(3	)	14			(4	)	112			(6	)	(1		)	(3	)	4			(3	)
Other comprehensive income (loss)	18					3					(17		)			(40		)			—					—					—					2					—
Purchases	4					—					35					189					—					—					—					—					—
Issuances	—					—					—					—					—					—					(68		)	(8	)	—					—
Settlements	(2		)			—					(64		)			(46		)			(116		)			(1		)			18					108					8
FG VIE deconsolidations	—					—					—					—					(17		)			—					—					1					16
Fair value as of December 31, 2018	$	99				$	56				$	309				$	947				$	569				$	77				$	(207	)			$	(517	)			$	(102	)
Change in unrealized gains/(losses) included in earnings related to financial instruments held as of December 31, 2018																					$	13		(3	)	$	14		(4	)	$	122		(6	)	$	1		(3	)	$	3		(3	)
Change in unrealized gains/(losses) included in OCI related to financial instruments held as of December 31, 2018	$	18				$	3				$	(14	)			$	(38	)								$	—									$	2

Fair Value Level 3 Rollforward

Recurring Basis

Year Ended December 31, 2017



	Fixed-Maturity Securities
	Obligations of State and Political Subdivisions					Corporate Securities				RMBS					Asset- Backed Securities					FG VIEs’ Assets at Fair Value					Other (7)					Credit Derivative Asset (Liability), net (5)					FG VIEs’ Liabilities with Recourse, at Fair Value					FG VIEs’ Liabilities without Recourse, at Fair Value
	(in millions)
Fair value as of December 31, 2016	$	39				$	60			$	365				$	805				$	876				$	65				$	(389	)			$	(807	)			$	(151	)
MBIA UK Acquisition	—					—				—					7					—					—					—					—					—
Total pretax realized and unrealized gains/(losses) recorded in: (1)
Net income (loss)	(13		)	(2	)	6		(2	)	27			(2	)	113			(2	)	37			(3	)	(2		)	(4	)	107			(6	)	(16		)	(3	)	(6		)	(3	)
Other comprehensive income (loss)	(2		)			1				23					56					—					—					—					—					—
Purchases	—					—				42					173					—					1					—					—					—
Settlements	(2		)			—				(123		)			(367		)			(147		)			—					13					145					12
FG VIE consolidations	—					—				—					—					39					—					—					—					(39		)
FG VIE deconsolidations	—					—				—					—					(105		)			—					—					51					54
Transfers into Level 3	54					—				—					—					—					—					—					—					—
Fair value as of December 31, 2017	$	76				$	67			$	334				$	787				$	700				$	64				$	(269	)			$	(627	)			$	(130	)
Change in unrealized gains/(losses) related to financial instruments held as of December 31, 2017	$	(2	)			$	1			$	23				$	123				$	59		(3	)	$	(2	)	(4	)	$	96		(6	)	$	(11	)	(3	)	$	(6	)	(3	)

____________________


(1)	Realized and unrealized gains (losses) from changes in values of Level 3 financial instruments represent gains (losses) from changes in values of those financial instruments only for the periods in which the instruments were classified as Level 3.


(2)	Included in net realized investment gains (losses) and net investment income.


(3)	Included in fair value gains (losses) on FG VIEs.


(4)	Recorded in net investment income and other income.


(5)	Represents the net position of credit derivatives. Credit derivative assets (recorded in other assets) and credit derivative liabilities (presented as a separate line item) are shown gross in the consolidated balance sheet based on net exposure by counterparty.


(6)	Reported in net change in fair value of credit derivatives.


(7)	Includes short-term investments, CCS and other invested assets.

(8) Relates to SGI Transaction. See Note 2, Assumption of Insured Portfolio and Business Combinations.

Level 3 Fair Value Disclosures

Quantitative Information About Level 3 Fair Value Inputs

At December 31, 2018



Financial Instrument Description(1)	Fair Value at December 31, 2018 (in millions)			Significant Unobservable Inputs	Range				Weighted Average as a Percentage of Current Par Outstanding
Assets (2):
Fixed-maturity securities:
Obligations of state and political subdivisions	$	99		Yield	4.5	%	-	32.7%	12.0%

Corporate securities	56			Yield	29.5%

RMBS	309			CPR	3.4	%	-	19.4%	6.2%
				CDR	1.5	%	-	6.9%	5.2%
				Loss severity	40.0	%	-	125.0%	82.7%
				Yield	5.3	%	-	8.1%	6.3%
Asset-backed securities:
Triple-X life insurance transactions	620			Yield	6.5	%	-	7.1%	6.8%

Collateralized loan obligations (CLOs)/TruPS	274			Yield	3.8	%	-	4.7%	4.3%

Others	53			Yield	11.5%

FG VIEs’ assets, at fair value	569			CPR	0.9	%	-	18.1%	9.3%
				CDR	1.3	%	-	23.7%	5.1%
				Loss severity	60.0	%	-	100.0%	79.8%
				Yield	5.0	%	-	10.2%	7.1%

Other assets	74			Implied Yield	6.6	%	-	7.2%	6.9%
Other assets	74			Term (years)	10 years
Liabilities:
Credit derivative liabilities, net	(207		)	Year 1 loss estimates	0.0	%	-	66.0%	2.2%
				Hedge cost (in basis points (bps))	5.5		-	82.5	23.3
				Bank profit (in bps)	7.2		-	509.9	77.3
				Internal floor (in bps)	8.8		-	30.0	19.0
				Internal credit rating	AAA		-	CCC	AA-

FG VIEs’ liabilities, at fair value	(619		)	CPR	0.9	%	-	18.1%	9.3%
				CDR	1.3	%	-	23.7%	5.1%
				Loss severity	60.0	%	-	100.0%	79.8%
				Yield	5.0	%	-	10.2%	5.6%

____________________


(1)	Discounted cash flow is used as the primary valuation technique for all financial instruments listed in this table.


(2)	Excludes several investments recorded in other invested assets with fair value of $7 million.

Quantitative Information About Level 3 Fair Value Inputs

At December 31, 2017



Financial Instrument Description(1)	Fair Value at December 31, 2017 (in millions)			Significant Unobservable Inputs	Range				Weighted Average as a Percentage of Current Par Outstanding
Assets (2):
Fixed-maturity securities :
Obligations of state and political subdivisions	$	76		Yield	4.5	%	-	40.8%	12.5%

Corporate securities	67			Yield	22.5%

RMBS	334			CPR	1.3	%	-	17.4%	6.4%
				CDR	1.5	%	-	9.2%	5.9%
				Loss severity	40.0	%	-	125.0%	82.5%
				Yield	4.0	%	-	7.5%	5.6%
Asset-backed securities:
Triple-X life insurance transactions	613			Yield	6.2	%	-	6.4%	6.3%

CLO/TruPS	116			Yield	2.6	%	-	4.6%	3.3%

Others	58			Yield	10.7%

FG VIEs’ assets, at fair value	700			CPR	3.0	%	-	14.9%	9.5%
				CDR	1.3	%	-	21.7%	5.4%
				Loss severity	60.0	%	-	100.0%	79.6%
				Yield	3.7	%	-	10.0%	6.2%

Other assets	60			Implied Yield	5.2	%	-	5.9%	5.5%
Other assets				Term (years)	10 years
Liabilities:
Credit derivative liabilities, net	(269		)	Year 1 loss estimates	0.0	%	-	42.0%	3.3%
				Hedge cost (in bps)	17.6		-	122.6	48.1
				Bank profit (in bps)	6.0		-	852.5	107.5
				Internal floor (in bps)	8.0		-	30.0	21.8
				Internal credit rating	AAA		-	CCC	AA-

FG VIEs’ liabilities, at fair value	(757		)	CPR	3.0	%	-	14.9%	9.5%
				CDR	1.3	%	-	21.7%	5.4%
				Loss severity	60.0	%	-	100.0%	79.6%
				Yield	3.4	%	-	10.0%	4.9%

____________________


(1)	Discounted cash flow is used as the primary valuation technique for all financial instruments listed in this table.


(2)	Excludes short-term investments with fair value of $1 million and several investments recorded in other invested assets with fair value of $7 million.

Not Carried at Fair Value

Financial Guaranty Insurance Contracts

For financial guaranty insurance contracts that are acquired in a business combination, the Company measures each contract at fair value on the date of acquisition, and then follows insurance accounting guidance on a recurring basis thereafter. In addition, the Company discloses the fair value of its outstanding financial guaranty insurance contracts. In both cases, fair value is based on management’s estimate of what a similarly rated financial guaranty insurance company would demand to acquire the Company’s in-force book of financial guaranty insurance business. It is based on a variety of factors that may include pricing assumptions management has observed for portfolio transfers, commutations, and acquisitions that have occurred in the financial guaranty market, as well as prices observed in the credit derivative market with an adjustment for illiquidity so that the terms would be similar to a financial guaranty insurance contract, and also includes adjustments to the carrying value of unearned premium reserve for stressed losses, ceding commissions and return on capital. The Company classified this fair value measurement as Level 3.

Long-Term Debt

Long-term debt issued by AGUS and AGMH is valued by broker-dealers using third party independent pricing sources and standard market conventions. The market conventions utilize market quotations, market transactions for the Company’s comparable instruments, and to a lesser extent, similar instruments in the broader insurance industry. The fair value measurement was classified as Level 2 in the fair value hierarchy.

The fair value of notes payable issued by AGM was determined by calculating the present value of the expected cash flows. The fair value measurement was classified as Level 3 in the fair value hierarchy.

The carrying amount and estimated fair value of the Company’s financial instruments not carried at fair value are presented in the following table.

Fair Value of Financial Instruments Not Carried at Fair Value



	As of December 31, 2018				As of December 31, 2017
	Carrying Amount		Estimated Fair Value		Carrying Amount		Estimated Fair Value
	(in millions)
Assets:
Other invested assets	$	1	$	2	$	8	$	9
Other assets (2)	130		130		97		97
Liabilities:
Financial guaranty insurance contracts (1)	3,240		5,932		3,330		7,104
Long-term debt	1,233		1,496		1,292		1,627
Other liabilities (2)	12		12		55		55

____________________


(1)	Carrying amount includes the assets and liabilities related to financial guaranty insurance contract premiums, losses, and salvage and subrogation and other recoverables net of reinsurance.


(2)	The Company’s other assets and other liabilities consist predominantly of accrued interest, receivables for securities sold and payables for securities purchased, for which the carrying value approximates fair value.