管理層發言
Greetings, and welcome to the Redwood Trust, Inc. Second Quarter 2026 Financial Results Conference Call. As a reminder, this conference is being recorded. It is now my pleasure to introduce your host, Natasha Fatheree, FP&A Leader. Thank you. You may begin.
Thank you, operator. Hello, everyone, and thank you for joining us today for Redwood's second quarter 2026 earnings conference call. With me on today's call are Chris Abate, Chief Executive Officer; Dash Robinson, President; Brooke Carillo, Chief Financial Officer; and Abhinav Asthana, our Chief Technology Officer. Before we begin today, I want to remind you that certain statements made during management's presentation today with respect to future financial and business performance may constitute forward-looking statements. Forward-looking statements are based on current expectations, forecasts and assumptions, which include risks and uncertainties that could cause actual results to differ materially. We encourage you to read the company's annual report on Form 10-K and quarterly report on Form 10-Q, which provide a description of some of the factors that could have a material impact on the company's performance and cause actual results to differ from those that may be expressed in forward-looking statements. On this call, we may also refer to both GAAP and non-GAAP financial measures. The non-GAAP financial measures provided should not be utilized in isolation or considered as a substitute for measures of financial performance prepared in accordance with GAAP. Reconciliations between GAAP and non-GAAP financial measures are provided in our second quarter Redwood review, which is available on our website, redwoodtrust.com. Also note that the contents of today's conference call contain time-sensitive information that are accurate only as of today. We do not intend and undertake no obligation to update this information to reflect subsequent events or circumstances. Finally, today's call is being recorded. It will be available on our website later today. With that, I'll turn the call over to Chris for opening remarks.
Thank you, and good morning, everyone. Redwood exceeded $8 billion of mortgage banking volume for the second straight quarter. We did over 20 securitizations in the first half of the year. We ended the quarter pricing three securitizations in a single week, one for each of our operating platforms, the first time for Redwood in our 32-year history. That makes us happy and a little nostalgic about how productive the company operates these days relative to the past when two to four securitizations a year was deemed just fine by market standards. Broadly speaking, it's no secret the housing finance business has been a lot less forgiving for this current generation of mortgage practitioners, first in over 40 years not to benefit from a long-term bull market in interest rates, which served as an invisible tailwind for both the lucky and the smart. Home affordability and supply headwinds, both closely linked to high interest rates and regulation, have impacted the addressable mortgage market and how mortgage businesses fundamentally operate. Today's environment requires higher operating efficiency and capital turnover and a deep strategic moat that can drive growth despite home sales activity still coming in at multi-decade lows. As investors seek to align with the long-term winners of this extended rate cycle, we're prioritizing a few key differentiators that are worth mentioning. Let's start with technology. We are rebuilding Redwood as an AI-native housing finance platform with proprietary systems developed by our own engineers and embedded directly into our workflows. Our multi-agent AI systems help teams retrieve answers quickly and apply the same intelligence to complex tasks, including seller financial reviews, guideline comparisons and contract analysis. The result has been faster expert reviews, greater consistency, and greater scale. There are people in the loop on every key decision. This is still early innings, but the capabilities we are deploying are proprietary, compounding and changing how we operate. Early indications of the operating leverage from technology are already visible. Direct expenses were 64 basis points as a percentage of volume for the first half of 2026, already a 28% improvement from full year 2025. Annualized time savings from our 2026 AI-enabled automation initiatives increased to approximately 23,600 hours, up more than 50% from the first quarter of 2026 baseline, with meaningful impacts on due diligence costs, rate sheet pricing and guideline analysis. We also extended our unified technology platform supporting Sequoia and Aspire to enable HELOCs as a new Sequoia product. The bottom line is this: if you're wondering who the AI winners and losers are going to be in housing finance, we'll put 90% annual volume growth with consistent margins up against anyone operating in the housing market today, a market that has been operating at overall volumes down 50% from 2021 levels. As many of you know, our RWT Horizons venture fund complemented, and in certain ways significantly accelerated, our growth in mortgage banking in recent years. Representing less than 2% of our capital, Horizons gives us access to more than 25 early-stage companies across the mortgage and AI ecosystem. During the quarter, we invested in Prometheus, an artificial intelligence company developing an artificial general engineer, while another AI company in our portfolio priced a financing round that values our initial seed investment at approximately 27x our cost. Our dual approach of adopting AI inside Redwood and investing directly at the frontier of technology remains a long-term strategic initiative. Product depth and distribution are another important part of the story. At Sequoia, newly launched products now represent more than 30% of our quarterly lock volume. Aspire also grew more than 30% sequentially in the non-QM space, while CoreVest is building momentum in its smaller balance offerings for experienced housing investors. Taken together, Redwood today is materially less dependent on any one product or on any mortgage refi cycle. It also differentiates our earnings model in comparison to monoline operators with revenues more tied to MSR values and associated customer retention. Our model conversely is built around efficiently aggregating loans from across our broad network and distributing them to long-term investors through securitizations, whole loan sales and strategic partnerships. Our bank relationships further strengthen that position. Large depositories leaned into mortgage volume during the second quarter, even at the expense of margins, underscoring that bank behavior is already evolving as the Basel III Endgame is finalized. Lower capital charges and high-quality mortgages may have been a necessary regulatory impediment for banks to reengage, but they are certainly not the only constraint. The ultimate decision by banks to boost origination activity remains risk-based. And to repeat ourselves, the mortgage risk that bank C-suites most consistently cite to us as top of mind is convexity, not credit. Redwood enables our bank partners to generate fee income and retain their clients while transferring their interest rate exposure to us while they retain and continue to grow the customer relationship. At June 30th, Redwood acted as a dedicated capital partner to 70% of the top 50 banks in the United States. Our ability to help banks manage ongoing mortgage exposures differentiates Redwood and reinforces our essential role throughout the banking system. In summary, the business we operate today is fundamentally different than it was 20, 10, or even two years ago. Advanced technology and operating efficiency, more comprehensive products, diversified distribution, premier institutional capital partnerships and a shrinking legacy portfolio position us to grow going forward through a wide range of market environments to create long-term value for shareholders, not just when all boats are rising, as they do when interest rates fall, but through challenging rate cycles where hard work and innovation make the difference. With that, I'll turn the call over to Dash to discuss our operating results.
Thank you, Chris. Our second quarter operating performance reflected the combined benefits of product diversification, capital-efficient distribution channels and an operating framework that's fully integrated with core AI initiatives at the center of our strategic blueprint. The result was an eighth consecutive quarter of mortgage banking returns north of 20%, increasingly fertile ground for continued capital redeployment away from our non-core portfolio holdings. At Sequoia, second quarter lock volume totaled $5.6 billion alongside several noteworthy product and distribution benchmarks. Gain on sale margins were 92 basis points overall, in line with the first quarter's 96 basis points, despite substantial macro headwinds in April and May and broader indications of pronounced margin compression across the industry. Distribution remained well-aligned with production, most notably with a Castlelake joint venture coming online in late June, nine Sequoia securitizations and $1.2 billion of whole loan sales, almost all to banks. Sequoia's production mix included over 65% purchase money loans. The strategic positioning Chris referenced has emerged as an important buffer against profitability headwinds for non-bank operators that are often coupled with reduced housing activity and renewed vigor for bank portfolios. This is in large part attributable to how our platform as a non-bank has positioned itself within the depository ecosystem. When business drivers, including those influenced by capital rules, need a bank to buy or sell mortgage loans, we are most often the first call. That deep bank relationship drove the launch of our medical professional loan program, now offered broadly to our seller network with great early success, including a second med pro securitization earlier in July that priced well inside our inaugural issuance. The recent launch of our HELOC program builds on our optimism that deeper product offerings will continue to drive resilience during periods of upward pressure on rates and volatility through stable margins, increased relevance to our deep seller network, and our ability to support two-way flow between bank portfolios. Also key to this positioning is Aspire, whose establishment 18 short months ago was designed to leverage existing strengths by offering a well-underwritten, flexible suite of expanded products to a broader network of originators. Aspire delivered over $2 billion of lock volume during the second quarter, another record for the platform, up 31% from Q1. Market observers expect non-QM originations to reach $150 billion in 2026, up 20% from last year and reflective of a growing cohort of high-quality borrowers that access credit differently than the traditional W-2 employee. This implies a run rate market share for Aspire of approximately 5% to 6% that we seek to grow to 10% by year-end 2026 through a relentless commitment to product innovation, accretive distribution and technology, including recently announced progress with AI-powered pricing and guideline analysis tools. Institutional investor demand continues to support the non-QM sector's growth in general and Aspire's in specific. The business completed its second and third securitizations issued under the Aspire shelf during the second quarter, with the risk retention and support in the tranches once again syndicated profitably to third-party investors. At June 30th, 60-plus day delinquencies within Aspire's securitized population were less than 10 basis points. Subsequent to quarter end, we executed definitive documentation for an Aspire-dedicated joint venture with Crayhill Capital Management, a leading structured credit investor. Through time, the vehicle has the potential purchasing power of up to $8 billion of loans, underscoring demand for Aspire's products and an important early validation for the business. Similar to our other joint ventures, it provides a source of recurring revenues with added performance fees upon reaching stated return thresholds. Each of our platforms now operates with a dedicated joint venture with key benefits to our operating leverage and revenue durability going forward. CoreVest, our direct originator focused on lending to housing investors, funded $410 million of loans during the second quarter, down approximately 5% from Q1 as higher rates weighed on portions of the pipeline and legislative uncertainty, now largely settled, impacted certain key pockets of market activity. We remain disciplined while borrowers and developers assess the evolving regulatory and legislative landscape. With the landmark housing bill now passed and build-for-rent carved out from institutional ownership limitations, activity is beginning to reopen in areas that had largely paused. CoreVest remains well-positioned, supported by its longstanding focus on experienced sponsors below the largest institutional segment. A key milestone for CoreVest during the quarter was its first term loan securitization since 2023, since which time our term loan production has largely been sold in whole loan form. The $268 million transaction priced accretively to loan sale economics and was placed with close to two dozen discrete investors, a market response that underscores the deep demand for the platform's origination activities. The team also entered into a new servicing arrangement later in the second quarter designed to reduce administrative demands and lower servicing costs over time and launched a targeted business development initiative to expand lead generation. As immediately realizable returns in mortgage banking continue to sit well above 20%, the value of continued reallocation away from our legacy investment segment remains significant. At quarter end, allocation to this portfolio totaled 12% of overall capital, down from 15% on March 31st and 63% lower than one year ago, when we announced the accelerated wind down of this position. Early in the third quarter, we commenced formal marketing of a substantial portion of our remaining legacy bridge loans and continued to progress individual line items through to resolutions, unlocking capital and reducing associated secured debt. Thus far in the third quarter, we also priced a new financing arrangement for the remainder of our home equity investment portfolio that pro forma we expect to reduce segment capital to below 10%. Ninety-day-plus delinquencies in the unsecuritized legacy bridge portfolio were roughly flat versus March 31. The priority remains fully moving on from this position as quickly and efficiently as possible to support further growth of our core activities. I will now turn the call over to Brooke to discuss our financial results.
Thank you, Dash. Turning to our second quarter results, we reported a GAAP net loss of $3 million, or $0.03 per share, compared with a $0.07 per share loss in the first quarter. Book value per common share was $6.90 at June 30. The 3% decline from $7.12 at March 31 was primarily driven by marked-to-market changes and ongoing carry costs within our legacy investments portfolio as well as the $0.18 dividend paid to common shareholders. On a non-GAAP basis, consolidated earnings available for distribution, or EAD, was $20 million or $0.15 per share, compared to $0.21 per share in the first quarter. The quarter again reflected two distinct trends. Our core segments remained highly profitable, generating $34 million of earnings available for distribution, representing an 18.5% annualized ROE, while legacy investments generated a $14 million EAD loss. Turning to our segment results: aggregate mortgage banking net revenue remained essentially flat despite a roughly 6% decline in production, reflecting stable to improving margins across the platforms while direct expenses declined. The result was a 33% annualized return on average capital for our operating platforms, with capital efficiency continuing to improve. Average capital required per dollar of production fell to roughly 2.6% in the first half of 2026 from about 3% a year ago, underscoring the scalability of our mortgage banking platforms as volumes grow. Prior to corporate allocations, Sequoia generated $32 million of GAAP net income compared with $38 million in the first quarter. The sequential decline was primarily volume driven as purchase commitments declined 9%, while the 92 basis point gain on sale margin remained near the high end of our historical target range. Cost per loan improved to 17 basis points from 18 basis points, demonstrating that we maintained operating discipline as volumes moderated. Initial loan transfers to Castlelake occurred near quarter end. Therefore, we expect the partnership to begin affecting capital velocity and fee economics more visibly in the second half of the year. Aspire generated $7 million of GAAP net income, up $5 million sequentially. Lock volume increased 31% to a record $2.1 billion, while gain on sale margins increased to 101 basis points from 73 basis points as securitization spreads normalized and hedge performance improved relative to the first quarter. Importantly, this growth was achieved with improving capital efficiency, resulting in a 33% annualized return on capital for the segment. CoreVest generated $1 million of GAAP net income, compared with a $3 million loss in the first quarter, which had included approximately $5 million of restructuring charges. Excluding acquisition-related expenses, EAD contribution for the segment increased to $3 million. Net revenue rose 8%, reflecting improved term loan execution, while direct operating expense declined meaningfully following the actions taken earlier this year. Net cost to originate was 96 basis points in the second quarter, up from 79 basis points in the first quarter, reflecting modestly lower fee and income relative to expenses, along with 5% lower quarter-over-quarter volume. Redwood Investments generated approximately $1 million of GAAP net income, compared with an $8 million loss in the first quarter. The improvement reflected a more constructive valuation backdrop across portions of the retained portfolio and lower expenses, although the segment continued to experience fair value pressure in selected bridge and SFR investments. We deployed $72 million of capital into investments sourced from second quarter securitizations. Because much of that deployment occurred late in the quarter, its earnings contribution should be more impactful in the third quarter. During the second quarter, we refinanced a portfolio of retained securities at an all-in cost of funds approximately 150 basis points below the prior financing. With approximately $1.5 billion of secured portfolio debt callable over the next 12 months, we retain meaningful optionality to reduce funding costs as opportunities arise. Legacy investments generated a $23 million GAAP loss, which included $12 million of negative fair value changes, primarily on legacy bridge loans inclusive of realized resolution activity. The financing, marketing and structured sale initiatives Dash discussed are intended to release capital for higher returning uses and reduce the negative carry still embedded in consolidated EAD. Based on the current return differential between legacy and our core segments, we estimate that each $100 million of capital successfully redeployed could improve consolidated EAD ROE by approximately 200 to 400 basis points through reinvestment in our operating platforms or potentially share repurchases at appropriate levels. Total operating expenses were down 21% on the quarter, with G&A declining to $38 million from $49 million. Approximately $7 million of the reduction reflected restructuring charges recorded in the first quarter, with the remainder primarily attributable to lower compensation and variable expenses. More importantly, first half adjusted expenses represented 64 basis points of production, compared with 88 basis points for the full year 2025 as volume growth continues to outpace expense growth. We expect some natural variability in quarterly expenses; the structural efficiency gains reflected in cost per loan trends and expenses relative to volume remain intact. Recourse debt declined by approximately $150 million to $4.5 billion, while recourse leverage declined modestly to 5x. More than half of recourse debt supports mortgage banking inventory that turns rapidly through securitizations or loan sales and joint ventures, with loans held for an average of approximately 26 days in June. We ended the quarter with $192 million of unrestricted cash, approximately $100 million of unencumbered assets and $3.7 billion of excess warehouse capacity. In the last year, we have renewed or added approximately $4.4 billion of capacity, and the senior notes issued in the quarter further extended our unsecured maturity profile. With that, I'll turn the call back to the operator for questions.
分析師問答
Our first question comes from Rick Shane with JPMorgan.
Can you hear me? Look, you guys are making progress in terms of reallocating capital. There's $195 million left. You're talking about getting down to 10% by the end of this quarter. Realistically, how much of that $195 million do you expect to be able to realize? Obviously, I think there's some friction as we saw this quarter. And as the business descales, there may be further operating losses associated with it. So how much of that $195 million melting actually will go into the remainder of the business over the next couple of years?
Rick, it's Dash. I can start. A couple of pieces to your question. We expect to continue trending the capital in the legacy investment segment to below 5% by the end of the year. That's how we've been guiding the market for a few quarters now. As we said in our prepared remarks, we actually did price a transaction this week, which we think pro forma will bring allocated capital to below 10% for that segment. That's definitely progress. As I also mentioned in the prepared remarks, we're currently working on a disposition plan for a large portion of the remaining unsecuritized bridge loans, which we'll hopefully have more to talk about for Q3 earnings. We believe we're still on track to have that segment below 5% of capital by the end of the year. As we said a lot, we're trying to be balanced between disposition speed and execution, also recognizing the significant accretion potential from redeployment of that capital. As we can elaborate on, we're highly confident that as that capital continues to come out of that segment, we will have a place to go with it immediately. We're still doing $8 billion-plus volumes in mortgage banking, and we're bringing on new joint ventures — all of which speak to the fact that those are tailwinds for us to continue to grow market share in mortgage banking. As Brooke articulated, the decisions around continuing to unlock that capital require weighing the right execution, but also the fact that there's approximately $0.14 to $0.15 a quarter of negative carry and opportunity cost within that segment that we think is immediately realizable through the retirement of secured debt, like I mentioned, and also the immediate redeployment. So we feel like the opportunities are there to redeploy very efficiently as we continue to wind that book down.
Got it. And how much — you executed a transaction at the beginning of the third quarter, as you've talked about. Presumably when you were valuing the portfolio at the end of the second, you were probably pretty close to that execution, so you had a good sense of value. How much of the second quarter mark was informed by the execution of the third quarter deal? Because again, I'm trying to understand, we saw capital allocation decline during the quarter, partially a portion of reallocation, but also partially a function of a decline of capital. That's what I'm trying to understand here: sort of that $195, how do we think about what flows into the rest of the business going forward?
Rick, I would say every asset in our legacy book at this point is distinct; we're down to a couple handfuls of loans. The execution we had in the third quarter is helpful, and we definitely were looking at what our resolution strategy was for each of the assets at June 30, and that informed our mark.
The transaction you're referring to, Rick, was for the remainder of our HEI position. Certainly, the mark at June 30 was informed by that execution, which we've since completed, so that's very much in line. As it relates to the legacy bridge portfolio, Brooke is right. As we say every quarter, that book is fair valued; it's marked where we feel like we could execute it. We'll be responsive to what the market tells us in terms of disposing of the rest, again with an eye toward where we can redeploy that capital quickly and reduce the secured debt that's influencing some of the carry costs Brooke articulated.
Our next question comes from Doug Harter with BTIG.
This is Will Nasta on for Doug this morning. I know you mentioned having a more cautious operating posture early in the quarter. Given the move higher in rates early this quarter, I was hoping you could talk about how you're thinking about banking volume sensitivity to rates and volatility versus higher rates, and how you guys are thinking about that right now. Also, you talked about your technology investment and how that's helped to improve expense efficiency down to 64 basis points. Where do you see that number trending? Do you see more potential upside there or progress you can make, or is there a particular level you are comfortable with?
We were more cautious in the second quarter. Earlier in the quarter, rates were very volatile and there was a lot of geopolitical uncertainty. June felt more stable and we leaned back in. Forty percent of our Q2 volume was in the month of June alone, which I think validates that we've got recurring revenue streams from these businesses and durable volume opportunities, but we're going to remain risk-minded. We saw things pick back up when we leaned in June, and we saw more of the same in July. In the past week or two, rates have backed up — the 10-year is around 4.63% and mortgage rates are near their one-year high. All of that we factor in. By and large, we feel pretty good about our risk position today and our ability to continue to grow volumes, but we can't control the macro environment and need to remain responsive. July has been a fairly strong month from a mortgage banking perspective, and we're hoping to maintain that momentum in August and September. This might be a good opportunity for Abhinav to chime in on some of the efficiencies we've been focused on, and then Brooke can follow up with some of the numbers.
Thank you, Dash. Redwood has been thoughtfully investing in technology and especially AI over the last 18 months. We've started to see that result in a compounding value proposition for the company. We've been investing in foundational AI platforms. As Chris mentioned, we're not bolting on AI with incremental, small changes. We are rethinking the operating model. As we built our platforms, we've re-engineered how our operating and business platforms conduct business. We've added efficiencies where we saw waste in the process, and we have eliminated parts of the function that no longer make sense. In doing so, we've provided value as we grow our businesses. More importantly, as we scale, these platforms are designed to handle volume and operate at efficiencies that will be significantly larger than where we are today.
The improvement thus far from 2025 has been driven first by the scalability of our platforms and the amount of market share we've gained, so volume has helped. Second, our variable expense structure has provided a large benefit, and we're starting to see technology carry some of its weight in the improvement. I think the next 10 to 15 basis points of improvement will probably be driven more by tech and continued scalability of our platform. We imagine this ratio will continue to decline as we efficiently fund our loans via technological enhancements that Chris, Abhinav and Dash mentioned in their prepared remarks.
Our next question comes from Marissa Lobo with UBS.
Just thinking about gain on sale margins, you flagged that banks were competing aggressively in Q2, but Sequoia margins were better than we expected. How much of that resilience was mix versus pricing discipline? And as banks lean in further, how should we think about how the gain on sale margins evolve?
We observed the large money center banks leaning back in, perhaps front-running the anticipated capital rule changes. That said, we've seen volume increases at meaningfully lower margins from some disclosed market participants, indicating some leaning in. We did a good job of maintaining our volumes while staying risk-minded, and staying risk-minded includes preserving margins and not chasing volume. Our business has been built to be a holistic partner to banks. In July, we locked a large bulk sale to a regional bank. We've been mostly buying loans from banks over the past few years, but there could be two-way flows. The essence of the franchise is the relationship itself and the technology implementations, LO training and other elements that go into a partnership. If banks want to lean in and want a capital partner, we're focused on serving our clients. That said, we don't necessarily see housing activity meaningfully higher and certainly refi activity trended down over the past quarter. These look like market share battles between banks and non-banks from an originator standpoint. We'll see, as Q2 earnings season completes, where overall volumes landed.
Can you provide any color on book value performance quarter to date?
We're up approximately 1% quarter to date. We've recovered part of Q2's decline.
That 1% is certainly a function of strong mortgage banking results and supply.
Our next question comes from Crispin Love with Piper Sandler.
What are your views on the administration focusing on housing, specifically housing affordability through GSE purchases, the single-family executive order, and related initiatives? Broadly, what do you think would be some of the best ways to address affordability issues in the U.S.?
The ROAD to Housing Act is very focused on housing supply, which is the right long-term answer. We need more homes built, easier permitting, and to ensure builders are profitable. There's a lot in the bill, and we were pleased build-to-rent wasn't adversely impacted. We're excited about the future of CoreVest. Supply initiatives are long-run solutions. In the short run, demand-side measures are what the administration can affect between now and the midterms. GSE MBS buying has been evident and has helped the TBA market through recent volatility. That has some offsetting pressures, which we suspect are coming from GSE purchases. Overall, that feeds into the non-agency space and has helped jumbo executions remain stable, which is positive. But in the near term, it is unclear what else can be done to meaningfully reduce mortgage rates. We are a long way from rates that would meaningfully boost refi volume. Home equity is a big initiative for many in the industry—ways to continue to serve clients without new mortgages. All of those things we're focused on. Between now and year-end, absent a major catalyst, we expect to remain range-bound.
On the ROAD to Housing legislation, we've seen CoreVest production a bit softer over the last two quarters largely tied to the legislation. Now that there's clarity, we have seen a pickup in transaction volume from middle-market investors, allowing them to reallocate capital. There's a lot of frozen capital on the sidelines, particularly in parts of the bridge market where we've been under-penetrated, especially in build-to-rent, which was about 2% of our volume this quarter. We might see a mix shift from some of that pent-up demand. Term sheets issued are up about 40% since the trough in the spring when the legislation was an overhang. CoreVest had a quarter where income picked up, and we should see more of that as some of these deals get done.
Our next question comes from Mikhail Goberman with Citizens JMP.
If I could get more color on your general thoughts on the non-QM space, what you are seeing in Aspire, your thoughts on the progression of lock volume going forward, which has been excellent, and your expectations for margins going forward.
Mikhail, we remain of the view that the non-QM market will continue to grow. As we said in the prepared remarks, there's expected growth of roughly 20% this year. We're leaning in at the right time. Consumer awareness has grown over the past couple of years, making consumers who qualify for these loans aware of their options. One of Aspire's value propositions has been the strong foundation from our Sequoia business and our years-long seller relationships, many of whom have begun to insource these expanded credit products to diversify product offerings and retain and attract originators. Two-thirds of Aspire production is with existing Sequoia relationships, which is close to our expectations. We are also adding new sellers and have many existing sellers not yet online. Growing to $2 billion a quarter is supported by that runway and underpins our goal to reach closer to 10% market share by year-end or early next year, up from an estimated 5% to 6% today. Regarding margins, we expect to be in our long-term range of 75 to 100 basis points. We're excited about the new Crayhill joint venture; like other JVs, it gives us pricing power and allows us to leverage internal capital 10x to 20x. The dollar goes further at higher ROEs when pairing our capital with third-party capital. It's become a virtuous cycle that has supported growth. The market continues to be responsive to these cash flows, and non-QM remains an efficient vehicle for investors to put capital to work in U.S. housing credit. We expect Aspire to continue to grow and for the market to absorb volumes.
If I could squeeze in one more, what are your general thoughts on borrower credit quality at mid-year?
In our experience, borrower credit quality has been quite stable. We track delinquencies and underwriting performance, and we've been fortunate with performance to this point. We're focused on working down our legacy book. Aspire and Sequoia performance has been consistent.
Our next question comes from Bose George with KBW.
I wanted to go back to the expenses discussion. Compensation expense was down quite a bit quarter over quarter. Was there some structural change, or was it that Q1 had some one-time items? Anything to call out there?
Thanks for asking. A portion of the Q1 number included about $5 million to $7 million of restructuring-related expenses that we expected to come out of our run rate. We had guided that we should be inside our fixed comp from Q4, which we saw in G&A by a couple of million dollars. About $7 million to $8 million was attributable to one-time items in Q1. We also had lower acquisition costs due to slightly lower volume, slightly lower portfolio management costs relative to Q1, and reductions in fixed comp and some variable costs. On an annualized basis, G&A is down probably $10 million to $12 million compared with that comparison point, while volume is up a couple of billion. That speaks to technology and scale. We're proud of those efficiency metrics.
Great. Makes sense. And then Chris, I didn't know if you mentioned this, but the allocation of capital looked like it reallocated from mortgage banking to the investment segment. Was that reflecting economics? What happened there?
Those shifts relate to servicing or other IO-related assets that hedge our pipeline. At a certain point, if those lose some of their pure hedging value for mortgage banking based on our pipeline, we will move them into the portfolio. We like those profiles as long-term hold assets as well. That was really the mix shift between the capital allocation between the portfolio and mortgage banking.
Okay. Was the decline in servicing income because of the reallocation?
We saw a slight pickup in speeds relative to our Q1 results. That was a small mark-to-market impact from legacy MSR.
We have reached the end of our question-and-answer session, which now concludes today's teleconference. You may disconnect your lines at this time. Thank you for your participation.