FAIR VALUE MEASUREMENT | FAIR VALUE MEASUREMENT U.S. GAAP requires the categorization of fair value measurement into three broad levels which form a hierarchy based on the transparency of inputs to the valuation. Level 1 - Quoted prices in active markets for identical instruments. Level 2 - Valuations based principally on other observable market parameters, including: • Quoted prices in active markets for similar instruments, • Quoted prices in less active or inactive markets for identical or similar instruments, • Other observable inputs (such as interest rates, yield curves, volatilities, prepayment rates, loss severities, credit risks and default rates), and • Market corroborated inputs (derived principally from or corroborated by observable market data). Level 3 - Valuations based significantly on unobservable inputs. New Residential follows this hierarchy for its fair value measurements. The classifications are based on the lowest level of input that is significant to the fair value measurement. The carrying values and fair values of New Residential’s assets and liabilities recorded at fair value on a recurring basis, as well as other financial instruments for which fair value is disclosed, as of December 31, 2016 were as follows: Fair Value Principal Balance or Notional Amount Carrying Value Level 1 Level 2 Level 3 Total Assets: Investments in: Excess mortgage servicing rights, at fair value (A) $ 277,975,997 $ 1,399,455 $ — $ — $ 1,399,455 $ 1,399,455 Excess mortgage servicing rights, equity method investees, at fair value (A) 60,677,300 194,788 — — 194,788 194,788 Mortgage servicing rights, at fair value (A) 79,935,302 659,483 — — 659,483 659,483 Servicer advances, at fair value 5,617,759 5,706,593 — — 5,706,593 5,706,593 Real estate securities, available-for-sale 8,788,957 5,073,858 — 1,530,298 3,543,560 5,073,858 Residential mortgage loans, held-for-investment 203,673 190,761 — — 190,343 190,343 Residential mortgage loans, held-for-sale 908,930 696,665 — — 717,985 717,985 Consumer loans, held-for-investment 1,809,952 1,799,486 — — 1,819,106 1,819,106 Derivative assets 6,776,052 6,762 — 6,762 — 6,762 Cash and cash equivalents 290,602 290,602 290,602 — — 290,602 Restricted cash 163,095 163,095 163,095 — — 163,095 Other assets 888,412 4,856 — — 4,856 4,856 $ 16,186,404 $ 453,697 $ 1,537,060 $ 14,236,169 $ 16,226,926 Liabilities: Repurchase agreements $ 5,193,686 $ 5,190,631 $ — $ 5,193,686 $ — $ 5,193,686 Notes and bonds payable 8,015,097 7,990,605 — — 7,993,326 7,993,326 Derivative liabilities 3,640,000 3,021 — 3,021 — 3,021 $ 13,184,257 $ — $ 5,196,707 $ 7,993,326 $ 13,190,033 (A) The notional amount represents the total unpaid principal balance of the residential mortgage loans underlying the MSRs and Excess MSRs. New Residential does not receive an excess mortgage servicing amount on non-performing loans in Agency portfolios. The carrying values and fair values of New Residential’s assets and liabilities recorded at fair value on a recurring basis, as well as other financial instruments for which fair value is disclosed, as of December 31, 2015 were as follows: Fair Value Principal Balance or Notional Amount Carrying Value Level 1 Level 2 Level 3 Total Assets Investments in: Excess mortgage servicing rights, at fair value (A) $ 329,367,971 $ 1,581,517 $ — $ — $ 1,581,517 $ 1,581,517 Excess mortgage servicing rights, equity method investees, at fair value (A) 73,058,050 217,221 — — 217,221 217,221 Servicer advances, at fair value 7,578,110 7,426,794 — — 7,426,794 7,426,794 Real estate securities, available-for-sale 4,418,552 2,501,881 — 917,598 1,584,283 2,501,881 Residential mortgage loans, held-for-investment 506,135 330,178 — — 330,433 330,433 Residential mortgage loans, held-for-sale 859,714 776,681 — — 784,750 784,750 Derivative assets 3,400,000 2,689 — 2,689 — 2,689 Cash and cash equivalents 249,936 249,936 249,936 — — 249,936 Restricted cash 94,702 94,702 94,702 — — 94,702 $ 13,181,599 $ 344,638 $ 920,287 $ 11,924,998 $ 13,189,923 Liabilities Repurchase agreements $ 4,043,942 $ 4,043,054 $ — $ 4,043,942 $ — $ 4,043,942 Notes and bonds payable 7,262,056 7,249,568 — — 7,260,909 7,260,909 Derivative liabilities 4,644,000 13,443 — 13,443 — 13,443 $ 11,306,065 $ — $ 4,057,385 $ 7,260,909 $ 11,318,294 (A) The notional amount represents the total unpaid principal balance of the residential mortgage loans underlying the Excess MSRs. New Residential does not receive an excess mortgage servicing amount on non-performing loans in Agency portfolios. New Residential has various processes and controls in place to ensure that fair value is reasonably estimated. With respect to the broker and pricing service quotations, to ensure these quotes represent a reasonable estimate of fair value, New Residential’s quarterly procedures include a comparison to quotations from different sources, outputs generated from its internal pricing models and transactions New Residential has completed with respect to these or similar assets or liabilities, as well as on its knowledge and experience of these markets. With respect to fair value estimates generated based on New Residential’s internal pricing models, New Residential corroborates the inputs and outputs of the internal pricing models by comparing them to available independent third party market parameters, where available, and models for reasonableness. New Residential believes its valuation methods and the assumptions used are appropriate and consistent with other market participants. Fair value measurements categorized within Level 3 are sensitive to changes in the assumptions or methodology used to determine fair value and such changes could result in a significant increase or decrease in the fair value. New Residential’s assets measured at fair value on a recurring basis using Level 3 inputs changed as follows: Level 3 Excess MSRs (A) Excess MSRs in Equity Method Investees (A)(B) Agency Non-Agency MSRs (A) Servicer Advances Non-Agency RMBS Total Balance at December 31, 2014 $ 217,519 $ 200,214 $ 330,876 $ — $ 3,270,839 $ 723,000 $ 4,742,448 Transfers (C) Transfers from Level 3 — — — — — — — Transfers to Level 3 — — — — — — — Transfers from investments in excess mortgage servicing rights, equity method investees, to investments in excess mortgage servicing rights — 98,258 (98,258 ) — — — — Gains (losses) included in net income Included in other-than-temporary impairment on securities (D) — — — — — (5,788 ) (5,788 ) Included in change in fair value of investments in excess mortgage servicing rights (D) (3,080 ) 41,723 — — — — 38,643 Included in change in fair value of investments in excess mortgage servicing rights, equity method investees (D) — — 31,160 — — — 31,160 Included in change in fair value of investments in servicer advances — — — — (57,491 ) — (57,491 ) Included in gain (loss) on settlement of investments, net — — — — — 3,061 3,061 Included in other income (loss), net (D) 2,852 147 — — — 879 3,878 Gains (losses) included in other comprehensive income (E) — — — — — (6,701 ) (6,701 ) Interest income 30,742 103,823 — — 352,316 69,632 556,513 Purchases, sales, repayments and transfers Purchases 254,149 917,078 — — 20,042,582 1,288,901 22,502,710 Proceeds from sales — — — — — (425,761 ) (425,761 ) Proceeds from repayments (64,981 ) (216,927 ) (46,557 ) — (16,181,452 ) (179,772 ) (16,689,689 ) Other transfers — — — — — 116,832 116,832 Balance at December 31, 2015 $ 437,201 $ 1,144,316 $ 217,221 $ — $ 7,426,794 $ 1,584,283 $ 10,809,815 Transfers (C) Transfers from Level 3 — — — — — — Transfers to Level 3 — — — — — — Gains (losses) included in net income Included in other-than-temporary impairment on securities (D) — — — — — (10,264 ) (10,264 ) Included in change in fair value of investments in excess mortgage servicing rights (D) (5,372 ) (1,925 ) — — — — (7,297 ) Included in change in fair value of investments in excess mortgage servicing rights, equity method investees (D) — — 16,526 — — — 16,526 Included in servicing revenue, net (F) 88,325 88,325 Included in change in fair value of investments in servicer advances — — — — (7,768 ) — (7,768 ) Included in gain (loss) on settlement of investments, net — — — — — (18,117 ) (18,117 ) Included in other income (loss), net (D) 2,452 350 — — — (4,875 ) (2,073 ) Gains (losses) included in other comprehensive income (E) — — — — — 124,669 124,669 Interest income 35,526 114,615 — — 364,350 209,706 724,197 Purchases, sales and repayments Purchases — 124 — 571,158 15,266,816 2,746,409 18,584,507 Proceeds from sales — — — — — (261,192 ) (261,192 ) Proceeds from repayments (88,050 ) (239,782 ) (38,959 ) — (17,343,599 ) (827,059 ) (18,537,449 ) Balance at December 31, 2016 $ 381,757 $ 1,017,698 $ 194,788 $ 659,483 $ 5,706,593 $ 3,543,560 $ 11,503,879 (A) Includes the recapture agreement for each respective pool. (B) Amounts represent New Residential’s portion of the Excess MSRs held by the respective joint ventures in which New Residential has a 50% interest. (C) Transfers are assumed to occur at the beginning of the respective period. (D) The gains (losses) recorded in earnings during the period are attributable to the change in unrealized gains (losses) relating to Level 3 assets still held at the reporting dates and realized gains (losses) recorded during the period. (E) These gains (losses) were included in net unrealized gain (loss) on securities in the Consolidated Statements of Comprehensive Income. (F) The components of Servicing revenue, net are disclosed in Note 5. Investments in Excess MSRs, Excess MSRs Equity Method Investees and MSRs Valuation Fair value estimates of New Residential’s MSRs and Excess MSRs were based on internal pricing models. The valuation technique is based on discounted cash flows. Significant inputs used in the valuations included expectations of prepayment rates, delinquency rates, recapture rates, the mortgage servicing amount or excess mortgage servicing amount of the underlying residential mortgage loans, as applicable, and discount rates that market participants would use in determining the fair values of mortgage servicing rights on similar pools of residential mortgage loans. In addition, for MSRs significant inputs included the market-level estimated cost of servicing. In order to evaluate the reasonableness of its fair value determinations, New Residential engages an independent valuation firm to separately measure the fair value of its MSRs and Excess MSRs. The independent valuation firm determines an estimated fair value range of each pool based on its own models and issues a “fairness opinion” with this range. New Residential compares the range included in the opinion to the value generated by its internal models. To date, New Residential has not made any significant valuation adjustments as a result of these fairness opinions. In addition, in valuing the MSRs and Excess MSRs, New Residential considered the likelihood of one of its servicers being removed as the servicer, which likelihood is considered to be remote. Significant increases (decreases) in the discount rates, prepayment or delinquency rates, or costs of servicing, in isolation would result in a significantly lower (higher) fair value measurement, whereas significant increases (decreases) in the recapture rates or mortgage servicing amount or excess mortgage servicing amount, as applicable, in isolation would result in a significantly higher (lower) fair value measurement. Generally, a change in the delinquency rate assumption is accompanied by a directionally similar change in the assumption used for the prepayment rate. The following tables summarize certain information regarding the weighted average inputs used in valuing the Excess MSRs, owned directly and through equity method investees: December 31, 2016 Significant Inputs (A) Prepayment Rate (B) Delinquency (C) Recapture Rate (D) Mortgage Servicing Amount (E) Collateral Weighted Average Maturity Years (F) Excess MSRs Directly Held (Note 4) Agency Original Pools 10.1 % 3.2 % 32.6 % 21 24 Recaptured Pools 7.4 % 4.3 % 23.0 % 21 25 Recapture Agreement 7.4 % 5.0 % 20.0 % 22 — 9.3 % 3.6 % 29.5 % 21 24 Non-Agency (G) Nationstar and SLS Serviced: Original Pools 11.8 % N/A 10.7 % 14 24 Recaptured Pools 7.9 % N/A 20.0 % 21 24 Recapture Agreement 7.5 % N/A 20.0 % 20 — Ocwen Serviced Pools 8.8 % N/A — % 14 26 9.4 % N/A 2.7 % 14 26 Total/Weighted Average--Excess MSRs Directly Held 9.4 % 3.6 % 10.0 % 16 26 Excess MSRs Held through Equity Method Investees (Note 4) Agency Original Pools 11.8 % 5.2 % 35.0 % 19 23 Recaptured Pools 7.3 % 4.5 % 24.7 % 23 25 Recapture Agreement 7.3 % 5.0 % 20.0 % 23 — Total/Weighted Average--Excess MSRs Held through Investees 9.8 % 5.0 % 29.8 % 21 24 Total/Weighted Average--Excess MSRs All Pools 9.5 % 3.9 % 14.2 % 17 26 MSRs Agency Ditech subserviced pools 12.7 % 3.2 % 29.1 % 26 23 FirstKey subserviced pools (H) 11.2 % 0.5 % 19.6 % 26 24 Total/Weighted Average--MSRs 12.4 % 2.8 % 27.5 % 26 23 December 31, 2015 Significant Inputs (A) Prepayment Rate (B) Delinquency (C) Recapture Rate (D) Excess Mortgage Servicing Amount (E) Collateral Weighted Average Maturity Years (F) Excess MSRs Directly Held (Note 4) Agency Original Pools 10.7 % 3.5 % 29.5 % 21 24 Recaptured Pools 7.5 % 4.9 % 20.0 % 20 25 Recapture Agreement 7.6 % 4.9 % 20.0 % 22 — 10.0 % 3.8 % 27.4 % 21 24 Non-Agency (G) Nationstar and SLS Serviced: Original Pools 12.5 % N/A 10.2 % 14 24 Recaptured Pools 7.5 % N/A 20.0 % 20 25 Recapture Agreement 7.5 % N/A 20.0 % 20 — Ocwen Serviced Pools 9.3 % N/A — % 14 26 10.0 % N/A 2.6 % 14 26 Total/Weighted Average--Excess MSRs Directly Held 10.0 % 3.8 % 9.5 % 16 25 Excess MSRs Held through Equity Method Investees (Note 4) Agency Original Pools 12.6 % 5.9 % 34.3 % 19 24 Recaptured Pools 7.7 % 5.0 % 20.0 % 23 25 Recapture Agreement 7.7 % 4.9 % 20.0 % 23 — Total/Weighted Average--Excess MSRs Held through Investees 10.8 % 5.6 % 29.0 % 20 24 Total/Weighted Average--Excess MSRs All Pools 10.2 % 4.2 % 13.6 % 17 25 (A) Weighted by fair value of the portfolio. (B) Projected annualized weighted average lifetime voluntary and involuntary prepayment rate using a prepayment vector. (C) Projected percentage of residential mortgage loans in the pool for which the borrower will miss its mortgage payments. (D) Percentage of voluntarily prepaid loans that are expected to be refinanced by the related servicer or subservicer, as applicable. (E) Weighted average total mortgage servicing amount, in excess of the basic fee as applicable, measured in basis points (bps). (F) Weighted average maturity of the underlying residential mortgage loans in the pool. (G) For certain pools, the Excess MSR will be paid on the total UPB of the mortgage portfolio (including both performing and delinquent loans until REO). For these pools, no delinquency assumption is used. (H) Recapture rate represents the expected recapture rate with the successor subservicer appointed by NRM. As of December 31, 2016 and 2015 , weighted average discount rates of 9.8% and 9.8% , respectively, were used to value New Residential’s investments in Excess MSRs (directly and through equity method investees). As of December 31, 2016 , a weighted average discount rate of 12.0% was used to value New Residential’s investments in MSRs. All of the assumptions listed have some degree of market observability, based on New Residential’s knowledge of the market, relationships with market participants, and use of common market data sources. New Residential uses assumptions that generate its best estimate of future cash flows for each investment in MSRs and Excess MSRs. When valuing MSRs and Excess MSRs, New Residential uses the following criteria to determine the significant inputs: • Prepayment Rate: Prepayment rate projections are in the form of a “vector” that varies over the expected life of the pool. The prepayment vector specifies the percentage of the collateral balance that is expected to prepay voluntarily (i.e., pay off) and involuntarily (i.e., default) at each point in the future. The prepayment vector is based on assumptions that reflect macroeconomic conditions and loan level factors such as the borrower’s interest rate, FICO score, loan-to-value ratio, debt-to-income ratio, vintage on a loan level basis, as well as the projected effect on loans eligible for the Home Affordable Refinance Program 2.0 (“HARP 2.0”). New Residential considers historical prepayment experience associated with the collateral when determining this vector and also reviews industry research on the prepayment experience of similar loan pools. This data is obtained from remittance reports, market data services and other market sources. • Delinquency Rates: For existing mortgage pools, delinquency rates are based on the recent pool-specific experience of loans that missed their latest mortgage payments. Delinquency rate projections are in the form of a “vector” that varies over the expected life of the pool. The delinquency vector specifies the percentage of the unpaid principal balance that is expected to be delinquent each month. The delinquency vector is based on assumptions that reflect macroeconomic conditions, the historical delinquency rates for the pools and the underlying borrower characteristics such as the FICO score and loan-to-value ratio. For the recapture agreements and recaptured loans, delinquency rates are based on the experience of similar loan pools originated by New Residential’s servicers and subservicers, and delinquency experience over the past year. New Residential believes this time period provides a reasonable sample for projecting future delinquency rates while taking into account current market conditions. Additional consideration is given to loans that are expected to become 30 or more days delinquent. • Recapture Rates: Recapture rates are based on actual average recapture rates experienced by New Residential’s servicers and subservicers on similar residential mortgage loan pools. Generally, New Residential looks to three to six months’ worth of actual recapture rates, which it believes provides a reasonable sample for projecting future recapture rates while taking into account current market conditions. Recapture rate projections are in the form of a “vector” that varies over the expected life of the pool. The recapture vector specifies the percentage of the refinanced loans that have been recaptured within the pool by the servicer or subservicer. The recapture vector takes into account the nature and timeline of the relationship between the borrowers in the pool and the servicer or subservicer, the customer retention programs offered by the servicer or subservicer and the historical recapture rates. • Mortgage Servicing Amount or Excess Mortgage Servicing Amount: For existing mortgage pools, mortgage servicing amount and excess mortgage servicing amount projections are based on the actual total mortgage servicing amount, in excess of a base fee as applicable. For loans expected to be refinanced by the related servicer or subservicer and subject to a recapture agreement, New Residential considers the mortgage servicing amount or excess mortgage servicing amount on loans recently originated by the related servicer over the past three months and other general market considerations. New Residential believes this time period provides a reasonable sample for projecting future mortgage servicing amounts and excess mortgage servicing amounts while taking into account current market conditions. • Discount Rate: The discount rates used by New Residential are derived from market data on pricing of mortgage servicing rights backed by similar collateral. New Residential uses different prepayment and delinquency assumptions in valuing the MSRs and Excess MSRs relating to the original loan pools, the recapture agreements and the MSRs and Excess MSRs relating to recaptured loans. The prepayment rate and delinquency rate assumptions differ because of differences in the collateral characteristics, eligibility for HARP 2.0 and expected borrower behavior for original loans and loans which have been refinanced. The assumptions for recapture and discount rates when valuing MSRs and Excess MSRs and recapture agreements are based on historical recapture experience and market pricing. Investments in Servicer Advances Valuation New Residential uses internal pricing models to estimate the future cash flows related to the Servicer Advance investments that incorporate significant unobservable inputs and include assumptions that are inherently subjective and imprecise. New Residential’s estimations of future cash flows include the combined cash flows of all of the components that comprise the Servicer Advance investments: existing advances, the requirement to purchase future advances, the recovery of advances and the right to the basic fee component of the related MSR. The factors that most significantly impact the fair value include (i) the rate at which the Servicer Advance balance changes over the term of the investment, (ii) the UPB of the underlying loans with respect to which New Residential has the obligation to make advances and owns the basic fee component of the related MSR which, in turn, is driven by prepayment rates and (iii) the percentage of delinquent loans with respect to which New Residential owns the basic fee component of the related MSR. The valuation technique is based on discounted cash flows. Significant inputs used in the valuations included the assumptions used to establish the aforementioned cash flows and discount rates that market participants would use in determining the fair values of Servicer Advances. In order to evaluate the reasonableness of its fair value determinations, New Residential engages an independent valuation firm to separately measure the fair value of its investment in Servicer Advances. The independent valuation firm determines an estimated fair value range based on its own models and issues a “fairness opinion” with this range. New Residential compares the range included in the opinion to the value generated by its internal models. To date, New Residential has not made any significant valuation adjustments as a result of these fairness opinions. In valuing the Servicer Advances, New Residential considered the likelihood of the related servicer being removed as the servicer, which likelihood is considered to be remote. Significant increases (decreases) in the advance balance-to-UPB ratio, prepayment rate, delinquency rate, or discount rate, in isolation, would result in a significantly lower (higher) fair value measurement. Generally, a change in the delinquency rate assumption is accompanied by a directionally similar change in the assumption used for the advance balance-to-UPB ratio. The following table summarizes certain information regarding the inputs used in valuing the Servicer Advances: Significant Inputs Weighted Average Outstanding Servicer Advances to UPB of Underlying Residential Mortgage Loans Prepayment Rate (A) Delinquency Mortgage Servicing Amount (B) Discount Rate Collateral Weighted Average Maturity (Years) (C) December 31, 2016 2.1 % 9.8 % 14.9 % 8.3 bps 5.6 % 24.8 December 31, 2015 2.3 % 10.4 % 17.5 % 9.2 bps 5.6 % 24.5 (A) Projected annual weighted average lifetime voluntary and involuntary prepayment rate using a prepayment vector. (B) Mortgage servicing amount excludes the amounts New Residential pays its servicers as a monthly servicing fee. (C) Weighted average maturity of the underlying residential mortgage loans in the pool. The valuation of the Servicer Advances also takes into account the performance fee paid to the servicer, which in the case of the Buyer is based on its equity returns and therefore is impacted by relevant financing assumptions such as loan-to-value ratio and interest rate, and which in the case of Servicer Advances acquired from HLSS is based partially on future LIBOR estimates. All of the assumptions listed have some degree of market observability, based on New Residential’s knowledge of the market, relationships with market participants, and use of common market data sources. The prepayment rate, the delinquency rate and the advance-to-UPB ratio projections are in the form of “curves” or “vectors” that vary over the expected life of the underlying mortgages and related Servicer Advances. New Residential uses assumptions that generate its best estimate of future cash flows for each investment in Servicer Advances, including the basic fee component of the related MSR. When valuing Servicer Advances, New Residential uses the following criteria to determine the significant inputs: • Servicer advance balance: Servicer advance balance projections are in the form of a “vector” that varies over the expected life of the residential mortgage loan pool. The servicer advance balance projection is based on assumptions that reflect factors such as the borrower’s expected delinquency status, the rate at which delinquent borrowers re-perform or become current again, servicer modification offer and acceptance rates, liquidation timelines and the servicers’ stop advance and clawback policies. • Prepayment Rate: Prepayment rate projections are in the form of a “vector” that varies over the expected life of the pool. The prepayment vector specifies the percentage of the collateral balance that is expected to prepay voluntarily (i.e., pay off) and involuntarily (i.e., default) at each point in the future. The prepayment vector is based on assumptions that reflect macroeconomic conditions and factors such as the borrower’s FICO score, loan-to-value ratio, debt-to-income ratio, and vintage on a loan level basis. New Residential considers collateral-specific prepayment experience when determining this vector. • Delinquency Rates: For existing mortgage pools, delinquency rates are based on the recent pool-specific experience of loans that missed recent mortgage payment(s) as well as loan- and borrower-specific characteristics such as the borrower’s FICO score, the loan-to-value ratio, debt-to-income ratio, occupancy status, loan documentation, payment history and previous loan modifications. New Residential believes the time period utilized provides a reasonable sample for projecting future delinquency rates while taking into account current market conditions. • Mortgage Servicing Amount: Mortgage servicing amounts are contractually determined on a pool-by-pool basis. New Residential projects the weighted average mortgage servicing amount based on its projections for prepayment rates. • LIBOR: The performance-based incentive fees on both Ocwen-serviced and Nationstar-serviced servicer advance portfolios are driven by LIBOR-based factors. The LIBOR curves used are widely used by market participants as reference rates for many financial instruments. • Discount Rate: The discount rates used by New Residential are derived from market data on pricing of mortgage servicing rights backed by similar collateral and the advances made thereon. Real Estate Securities Valuation New Residential’s securities valuation methodology and results are further detailed as follows: Fair Value Asset Type Outstanding Face Amount Amortized Cost Basis Multiple Quotes (A) Single Quote (B) Total Level December 31, 2016 Agency RMBS $ 1,486,739 $ 1,532,421 $ 1,530,298 $ — $ 1,530,298 2 Non-Agency RMBS (C) 7,302,218 3,415,906 3,028,094 515,466 3,543,560 3 Total $ 8,788,957 $ 4,948,327 $ 4,558,392 $ 515,466 $ 5,073,858 December 31, 2015 Agency RMBS $ 884,578 $ 918,633 $ 917,598 $ — $ 917,598 2 Non-Agency RMBS (C) 3,533,974 1,579,445 1,029,981 554,302 1,584,283 3 Total $ 4,418,552 $ 2,498,078 $ 1,947,579 $ 554,302 $ 2,501,881 (A) New Residential generally obtained pricing service quotations or broker quotations from two sources, one of which was generally the seller (the party that sold New Residential the security) for Non-Agency RMBS. New Residential evaluates quotes received and determines one as being most representative of fair value, and does not use an average of the quotes. Even if New Residential receives two or more quotes on a particular security that come from non-selling brokers or pricing services, it does not use an average because it believes using an actual quote more closely represents a transactable price for the security than an average level. Furthermore, in some cases there is a wide disparity between the quotes New Residential receives. New Residential believes using an average of the quotes in these cases would not represent the fair value of the asset. Based on New Residential’s own fair value analysis, it selects one of the quotes which is believed to more accurately reflect fair value. New Residential has not adjusted any of the quotes received in the periods presented. These quotations are generally received via email and contain disclaimers which state that they are “indicative” and not “actionable” — meaning that the party giving the quotation is not bound to actually purchase the security at the quoted price. New Residential’s investments in Agency RMBS are classified within Level 2 of the fair value hierarchy because the market for these securities is very active and market prices are readily observable. The third-party pricing services and brokers engaged by New Residential (collectively, “valuation providers”) use either the income approach or the market approach, or a combination of the two, in arriving at their estimated valuations of RMBS. Valuation providers using the market approach generally look at prices and other relevant information generated by market transactions involving identical or comparable assets. Valuation providers using the income approach create pricing models that generally incorporate such assumptions as discount rates, expected prepayment rates, expected default rates and expected loss severities. New Residential has reviewed the methodologies utilized by its valuation providers and has found them to be consistent with GAAP requirements. In addition to obtaining multiple quotations, when available, and reviewing the valuation methodologies of its valuation providers, New Residential creates its own internal pricing models for Level 3 securities and uses the outputs of these models as part of its process of evaluating the fair value estimates it receives from its valuation providers. These models incorporate the same types of assumptions as the models used by the valuation providers, but the assumptions are developed independently. These assumptions are regularly refined and updated at least quarterly by New Residential, and reviewed by its valuation group, which is separate from its investment acquisition and management group, to reflect market developments and actual performance. For 77.1% of New Residential’s Non-Agency RMBS, the ranges of assumptions used by New Residential’s valuation providers are summarized in the table below. The assumptions used by New Residential’s valuation providers with respect to the remainder of New Residential’s Non-Agency RMBS were not readily available. Fair Value Discount Rate Prepayment Rate (a) CDR (b) Loss Severity (c) Non-Agency RMBS $ 2,731,218 2.06% to 32.75% 0.25% to 20% 0.25% to 10.0% 5.0% to 100% (a) Represents the annualized rate of the prepayments as a percentage of the total principal balance of the pool. (b) Represents the annualized rate of the involuntary prepayments (defaults) as a percentage of the total principal balance of the pool. (c) Represents the expected amount of future realized losses resulting from the ultimate liquidation of a particular loan, expressed as the net amount of loss relative to the outstanding balance. (B) New Residential was unable to obtain quotations from more than one source on these securities. For approximately $509.6 million in 2016 and $228.5 million in 2015 , the one source was the party that sold New Residential the security. (C) Includes New Residential’s investments in interest-only notes for which the fair value option for financial instruments was elected. For New Residential’s investments in real estate securities categorized within Level 3 of the fair value hierarchy, the significant unobservable inputs include the discount rates, assumptions related to prepayments, default rates and loss severities. Significant increases (decreases) in any of the discount rates, default rates or loss severities in isolation would result in a significantly lower (higher) fair value measurement. The impact of changes in prepayment rates would have differing impacts on fair value, depending on the seniority of the investment. Generally, a change in the default assumption is accompanied by directionally similar change |