One Belvedere Place Suite 300 Mill Valley, CA 94941

July 22, 2015

VIA EDGAR AND E-MAIL

Securities and Exchange Commission
Division of Corporation Finance
100 F Street, N.W.
Washington, D.C. 20549

Branch Chief

Division of Corporation Finance

Responses to Comments on:

Form 10-K for the Fiscal Year Ended December 31, 2014

Filed on February 25, 2015

Form 10-Q for the Quarterly Period Ended March 31, 2015

Filed May 7, 2015

File No. 1-13759

Dear Mr. John,

On behalf of Redwood Trust, Inc. (“Redwood”), I hereby provide the following response in reply to the Staff’s comment letter dated June 24, 2015 (the “Comment Letter”) in connection with the above-referenced Annual Report on Form 10-K (the “2014 Form 10-K”) and Quarterly Report on Form 10-Q (the “2015 Q1 Form 10-Q”). For your convenience, each of my responses is preceded with an italicized recitation of the comment set forth in the Comment Letter.

Form 10-K for the fiscal year ended December 31, 2014

Item 7. Management’s Discussion and Analysis of Financial Condition and Results of Operations

We own MSRs associated with both jumbo and conforming residential mortgage loans, which we refer to as “Jumbo MSRs” and “Conforming MSRs,” respectively. Our MSRs are retained from the sale of loans or are purchased on a stand-alone basis, as outlined on page 63 of the 2014 Form 10-K.

Base and excess MSR

We distinguish base (or “basic”) and excess MSRs in accordance with IRS specified “safe harbor” levels of servicing fees they consider to be reasonable compensation (or “base” fees) for servicing various loan types. For conforming loans, the IRS considers fees up to 0.25% (of associated loan principal) to be base fees, and for jumbo loans, fees up to 0.375% (of associated loan principal) to be base fees.

Our Jumbo MSRs entitle us to a contractually specified servicing fee, with rates ranging from 0.25% to 0.375%, and are therefore all considered base fees under the IRS safe harbor. As of December 31, 2014 and 2013, the weighted average servicing fee rate on our Jumbo MSRs was 0.25%. Our Conforming MSRs entitle us to a contractually specified servicing fee, with rates ranging from 0.25% to 0.70%. As of December 31, 2014 and 2013, our portfolio of Conforming MSRs had a fair value of $81.3 million and $3.3 million, respectively, and of these amounts MSRs with fair values of approximately $100,000 and $30,000, respectively, had servicing fees in excess of 0.25%.

MSR Yields

Our gross cash yield on MSRs (calculated by dividing the annual gross servicing fees we received, by the weighted average notional balance of loans associated with MSRs we owned during the year) was 0.23%, 0.23%, and 0.18% for the years ended December 31, 2014, 2013 and 2012, respectively.

Servicer Advances

At both December 31, 2014 and December 31, 2013, we had approximately $1.0 million and $800,000, respectively, of servicer advances, primarily related to recoverable escrow advances, presented in “Other assets” on our balance sheet.

In accordance with the comment letter request, in future filings, we will update our disclosures to include the amount of MSRs we own with excess servicing and the amount of servicing advances associated with MSRs as of each balance sheet date presented, as well as the gross cash yield on our MSRs for each period presented in our statements of income.

Item 8. Financial Statements and Supplementary Data

We carry our jumbo residential mortgage loans (“jumbo loans”) at fair value, as they have historically represented our loan inventory for our residential mortgage banking activities. Our jumbo loans held-for-sale have typically been held on balance sheet from 30-60 days, until they are sold or securitized. With the reasonably high turnover, quarter-end estimates of fair value for these loans are quickly realized in subsequent quarters.

Since prices or quotes from exchanges or listed markets are not available for jumbo loans, we estimate fair value for these loans using internal models that incorporate various observable and unobservable inputs, including the transactional activity noted above. We have not viewed the various purchasers of jumbo loans (e.g., whole loan investors, resellers, or securitization aggregators) as representative of separate markets, but rather as part of a single “secondary market” for jumbo loans. In fact, many purchasers fall into more than one of these categories and acquire jumbo loans for differing reasons. Similarly, sellers of jumbo loans typically seek bids for jumbo loans from many different types of purchasers, rather than solely from one category of purchasers. We view this single secondary market as the principal market, with various market participants providing varying pricing inputs each quarter. During 2014, the difference in fair value estimates implied by pricing inputs provided by different types of purchasers was minimal.

In considering the Staff’s comment, we plan to update our disclosures in future filings to clarify the existence of a single principal market for jumbo loans, as opposed to two distinct markets. The updated language we intend to use is as follows:

Estimated fair values for residential loans are determined using models that incorporate various observable and unobservable inputs, including pricing information from recent securitizations and whole loan sales. Certain significant inputs in these models are considered unobservable and are therefore Level 3 in nature. Pricing inputs obtained from market securitization activity include indicative spreads to indexed TBA prices for senior RMBS and indexed swap rates for subordinate RMBS, which are adjusted as necessary for current market conditions (Level 3). Pricing inputs obtained from market whole loan transaction activity include indicative spreads to indexed swap rates, adjusted as necessary for current market conditions (Level 3). Other observable inputs include Agency RMBS pricing, indexed swap yields, credit rating agency guidance on expected credit support levels for newly issued RMBS transactions, benchmark interest rates, and prepayment rates. These assets would generally decrease in value based upon an increase in the credit spread, prepayment speed, or credit support assumptions.

Estimated fair values for conforming loans are determined based upon quoted market prices (Level 2). Conforming loans are mortgage loans that conform to Agency guidelines. As necessary, these values are adjusted for servicing value, market conditions and liquidity.

Form 10-Q for the quarterly period ended March 31, 2015

Item 2. Management’s Discussion and Analysis of Financial Condition and Results of Operations

We purchase jumbo residential mortgage loans (“jumbo loans”) from various bank and non-bank loan originators, which we refer to as “Sellers.” Our purchases of jumbo loans from these Sellers are governed by mortgage loan purchase and sale agreements (or “MLPSAs”). Prior to January 1, 2015, our MLPSAs were drafted such that there was no legally enforceable commitment by us to purchase a jumbo loan that we and the Seller had specified until a purchase price and terms letter (“PPTL”) relating to that loan was executed by both parties. Once the PPTL was executed by both parties, a contractual purchase and sale commitment between the parties was established; and, consequently, it was only at the time the PPTL was executed that a commitment to purchase a jumbo loan could be assessed under derivatives accounting guidance. Of note, this commitment does not represent an “Interest Rate Lock Commitment” to a borrower as we do not originate any residential loans ourselves.

Prior to January 1, 2015, we generally entered into PPTLs on the same day we purchased the related jumbo loan – i.e., on the same day we wired the purchase price to the Seller and the Seller conveyed ownership of the loan to us. Under this framework, even if an executed PPTL were to qualify as a derivative, we did not have open PPTLs at any quarter-end (because commitments to purchase jumbo loans were made and fulfilled on the same day) and, therefore, had no jumbo loan purchase commitments to assess as derivatives for financial reporting purposes.

During the latter part of 2014, we executed amendments to the MLPSAs we had in place with Sellers to affect certain new terms relating to purchase and sale commitments. Under the amendments, these new terms became effective on January 1, 2015. In addition, we changed our standard form MLPSA to affect the same new terms in new MLPSAs we entered into with new Sellers on and after January 1, 2015.

As of January 1, 2015, all of our MLPSAs specify that our commitment to purchase a jumbo loan (and the Seller’s corresponding commitment to sell us that loan) is established when we deliver a confirmation to the Seller relating to that loan. We now typically deliver a confirmation 30-45 days prior to when we expect to fulfill our commitment to purchase a loan. Because a contractual commitment is established well before a jumbo loan will be purchased, beginning with the quarter ended March 31, 2015, we assessed our open commitments to purchase jumbo loans under derivative accounting guidance to determine if these open commitments qualified as derivatives.

In analyzing these open commitments, we looked to ASC 815-10-15, paragraphs 69-71, which discuss the accounting treatment for “Certain Loan Commitments.” In accordance with paragraph 70 (formerly DIG C13), all commitments to purchase or sell mortgage loans must be evaluated under the definition of a derivative. Therefore, we have evaluated open commitments to purchase jumbo loans using the guidance in ASC 815-10-15-83, “Derivatives and Hedging – Definition of Derivative Instrument.” In accordance with this guidance, we determined that our current MLPSAs and associated confirmations are contractual commitments and evaluated the following required criteria to assess whether they meet the definition of a derivative:

With respect to our jumbo loans, the related MLPSA and confirmation evidence a purchase and sale obligation (a settlement requirement), specify the principal amount of the loan to be purchased, and specify the purchase price for the loan.

This satisfies the first criterion under ASC 815-10-15-83’s definition of a derivative.

With respect to our jumbo loans, the related MLPSA and confirmation require no initial net investment.

This satisfies the second criterion under ASC 815-10-15-83’s definition of a derivative.

ASC 815-10-15 paragraphs 99-139 discuss net settlement provisions. We evaluated each of the three means by which the net settlement criterion can be satisfied and determined that our underlying jumbo loans are readily convertible into cash.

This satisfies the third criterion under ASC 815-10-15-83’s definition of a derivative.

Accordingly, as we meet the specified criteria in ASC 815-10-15, we concluded that our current jumbo loan purchase commitments are considered derivatives in accordance with GAAP and we began to account for commitments entered into under our amended MLPSAs as derivatives beginning on January 1, 2015.

At March 31, 2015, we had $5.3 million of derivative assets and $0.8 million of derivative liabilities associated with jumbo loan purchase commitments recorded on our balance sheet. These amounts are included in our disclosures on page 37 of our 2015 Q1 Form 10-Q.

*      *      *

As you have requested, we confirm that:

Should you have any further comments or questions about this letter, please contact me by telephone at 415-384-3584, by fax at 415-381-1773, or by email at [email protected].