Rithm Property Trust Inc. (RPT) Q3 2017 Earnings Report, Transcript and Summary

Good afternoon and welcome to the Great Ajax Corp. Third Quarter 2017 Financial Results Conference Call. All participants will be listen-only mode. [Operator Instructions] Please note, this event is being recorded. I would now like to turn the conference over to Lawrence Mendelsohn, CEO. Please go ahead sir.

Thank you very much. Thank you everybody for joining us on our third quarter 2017 conference call. I'd like to, before we get started, just have everybody take a look at page two, the Safe Harbor Disclosure and then we can get on with the call. So, if we jump right to page three, in general a pretty good quarter. Good underlying economics of the loans we own and have bought, pretty exciting subsequent events as well, which we'll talk about more later, and some good NAV creation on the performance of our loans. A little more REO noise from foreclosed NPLs which I'll also talk about. We see and are experiencing continuing positive developments in loan markets and the securitization markets and our joint venture partnerships, developing those as well, and they are providing significant opportunities. In Q4, we have a lot going on as we'll see when we get to the subsequent event page. A little brief overview of Great Ajax Corp. for those a little bit less familiar, we use longstanding relationships to find loans from a diverse group of sellers. Our sourcing network is really, really, really important to what we do. It gives us the ability to acquire the types of loans that we want as opposed to just being an index fund, and it gives us the ability to buy them at the prices that we pay relative to others and we'll talk more about that a little later as well. We use our manager's proprietary analytics. We spend a lot of time analyzing data, large amounts of data for a variety of different reasons. Some of it to determine target loan characteristics and forecast performance patterns, some of it for geographic target market determination. The data analysis also helps us drive our loan servicing strategies with our captive servicer and give us daily updating expectations on loan-by-loan performance. Again, captive servicer, our affiliated servicer services loans asset-by-asset, borrower-by-borrower. Our servicer and our closest to our servicer is very important. I'll talk more about that later as well in subsequent events. It's not actually there, but it's more subsequent than even this presentation, so I'll talk about that as well. If we talk about leverage quickly, we use pretty moderate non-mark-to-market leverage. Our asset level leverage is 2.55 at quarter end. That actually is down from June 30. It was 2.63 at June 30, it's now 2.55. And our corporate leverage is 2.9 times. That's actually decreased from 2.91 at June 30 even though we issued $20.5 million of additional convertible notes in August, but corporate leverage still down slightly. In addition, we have a significant number of unlevered loans that we hold. Some highlights of Q3 ending September 30. We bought purchase price $26 million, about $32 million, $33 million of principal balance, and about $40 million, $41 million of underlying collateral. Eight transactions, purchase price continues to be pretty steady, 81.3% of UPB, 65% of property value. Keep in mind Q3 is generally a slower quarter from the acquisitions front than Q2 and Q4. You probably remember in Q2 we bought about $250 million of loans and in Q4, there's a lot of loans that we are in due diligence on as well. But from Q3, good purchases and we'll be able to do a lot with them. In August we issued $20.5 million as a follow-on of our convertible senior notes that we did in April. We issued them at 104% of par and as I mentioned earlier, total corporate leverage even with the additional $20.5 million of senior notes is actually lower at the end of Q3 than the end of Q2 and a lot of that has to do with performance of loans and pay down. Portfolio interest income, $24.5 million. The increase in net interest income is very much because of the Q2 loan purchases being on our balance sheet for the entire quarter, partially offset by additional interest expense from the convertible note issuance on the net interest income side. From a loan portfolio, we continue to see increases in reperformance and cash flow, dramatic cash flow. As a result, we have more cash flow and we have it for longer. It also makes the loans significantly more valuable on an NAV basis. We'll talk a little bit about that also later on. From an REO front, we had an additional GAAP write-down of approximately $1.1 million. About 30% of that came from one property in New Jersey that the owner purposely trashed. It was a higher end property in a nice part of New Jersey and during the eviction process, the owner purposely trashed the house. We discussed before that when we get back properties, frequently there is some adverse selection that can come from REO on the foreclosed NPL. One thing to keep in mind is that when we buy non-performing loans, a fair amount of them actually reperform and then a percentage continues to default and we have to foreclose. 43% of all the loans we foreclose on end up paying off either by refinancing a short sale or just selling the property or just getting a new loan. So, the loans that become REO are the 57% of the foreclosed property that don't do that. So, they are adversely selected in the sense that they are defaulted loans, they are a second time adversely selected in the sense that they are not one of the loans that reperformed and they are potentially adversely selected one more time because they are not part of the 43% that paid off because of the REO. The other thing you should think about is when a loan pays off in a foreclosure, it doesn't show up as an REO sale, it shows up as a loan payoff and gets in pool accounting for our loan portfolio so it gets blended into all our other loan cash flows. So, it's only loans that don't do that that become actual REO that could be subject to REO impairment. To the extent that properties are worth more than our basis, we don't get to write them up and that's just the nature of GAAP. If we didn't have the $1.1 million REO impairment, that cost us about $0.06 in our income, so basic earnings would have been $0.47 versus $0.41 for the REO. Another thing to keep in mind is that our NPL portfolio is pretty small relative to our total portfolio. Now, in late 2014, it was over 35% of our portfolio and now as you'll see, it's a little more than 4% of our portfolio. So, we're at the absolute peak of REO and REO selling that we're going to see because of all our NPL acquisitions in Q4 2014 and Q1 of 2015. Since the early part of 2015, we haven't bought many NPLs and as a result, you'll see REO creation slow down relative to what it's been the previous three quarters. Taxable income of $0.12 a share, I want to talk about that because that's very REO tax-driven as opposed to anything else. A tax gain or loss on REO is actually recognized at the time of foreclosure and it's equal to the difference between the tax basis and the fair value of the property. So, we're required to go get a valuation of the fair value of the property without regard to condition and without regard to selling costs. So, it's just what should this sell at if it was at fair value, good condition. That is the value that your tax basis gain is calculated by. No adjustments made for any liquidation costs, it generally results in a tax gain at the time of foreclosure. When you actually sell that REO, a corresponding tax loss is usually realized as the liquidation proceeds are reduced because you have selling costs like a broker commission, title expense. And you've also had servicer advances to make any repair or renovations. So, at the time of foreclosure, you have a tax gain and at the time of sale you frequently have a tax loss. Given that our REO is peaking and our REO sales are peaking, we take the tax loss when we sell those REO and we're selling a significant number of REO now versus previously because of our REO inventory from our Q1 2015 and Q4 2014 acquisitions. As a result of the peaking REO and peaking REO sales, this tax gets magnified and happens to be in this quarter. And that's just one of the natures of being a REIT. NPLs are not tax efficient in a REIT structure, one of the reasons among many that we don't particularly love NPLs for our portfolio and have dramatically decreased it, aside from the fact that we think that NPLs in today's market are overpriced. Book value, $15.60 at September 30, and also September 30, $43 million of cash on hand. We have plenty of cash and we have plenty of available credit, non-levered assets as well. Also, closing out the last day of September, another joint venture with DoubleLine Capital. We purchased $101 million -- we purchased loans for $101 million, about $110 million UPB. We paid about 59% of the underlying property value. This is the third joint venture closing with DoubleLine Capital and we expect that joint venture to continue in a material way. We're also in the documentation phase with DoubleLine on a small balance commercial venture that we expect to kick off the 1st of January. If we flip to page five, you'll see our portfolio and this is where you can really get a feel for how RPLs have taken over. Our RPL portfolio is now 95.5% of loans we own and NPL is about 4.5%. And from a property value perspective, it's the same combination about $33 million REO and rental properties. REO is really primarily held for sale and will turn into cash over a relatively short period of time. A small amount is held as rentals and those are generally multiunit properties, although we do have a rental property out in Bridgehampton, East of New York. On page six, you can see our RPLs. One of the trends that we're seeing for us and probably not necessarily the market, but for our RPLs, we continue to buy very low LTV loans with the overall RPL purchase price is 62.2% of the collateral value and 81.2% of UPB. We continue to playoff in some defense at the same time. The purchase price to collateral value actually declined in Q3 from 62.8% to 62.2% and it's down from 64.8% at 12/31. So, our purchase price relative to collateral value has actually come down nearly 3% this year which I don't think is typical of the loan markets in general. If we look at our NPL portfolio, purchased NPLs continue to decline in absolute dollars as well as a percentage of our portfolio. As previously described, this is on purpose. On our non-performing portfolio, the purchase price to collateral value is about 54.7% and the purchase price to UPB is about 62.5%. So we own them, the non-performing loans we own, we own pretty cheap. That being said, we're not eager to go out and buy additional non-performing loans. Our portfolio concentration remains pretty similar to last quarter, California still represents approximately 30% of our overall loan portfolio. And if you were to map that out, Santa Barbara, south to San Diego is probably 75% of that 30%. We've also purchased a number of small balance commercial loans, particularly small multifamily in Long Beach, Hawthorne, and Englewood areas in the LA Metro market. One thing that is not here is we've started to look at Houston again. We haven't added it as a key marketplace, but we're spending a lot of time analyzing Houston data and we've made five or six trips down to Houston pre and post-Hurricane Harvey. We've begun work also on several small balance commercial assets, multifamily in one particular neighborhood in Houston and we'll see how that progresses over the next couple of months. If we jump to page nine, portfolio migration is a very important topic for us. We've seen significant reperformance of our portfolio. If we look at the table and we just add up the 12 for 12 and the 24 for 24, that's about $600 million in UPB and if we add in 7 for 7, that's about $853 million of our portfolio. One of the things we find from the loans we buy, and I wouldn't say that this is all loans across the Board everywhere in the country, but for the particular loans we buy with their characteristics, once a loan gets 7 of 7, it reperforms a little -- it continues that reperformance through 12 of 12 a little over 90% of the time. And so we would expect that 12 for 12 number to increase pretty dramatically. Second on this page, I'd like to talk about any loan purchased after October 1 of 2016, which is less than 12 months ago, can't possibly be 12 for 12 or higher in payments to us because we haven't owned it for 12 months. So, the $598 million or the $600 million of 12 for 12 or better, doesn't include any loan that we haven't owned for at least a year which means the $280 million that we bought in Q2 and Q3 this year are not in those numbers. So, third, if we take into account payments made to prior servicers, so if we say, okay, let's make this table not just payments to us, but also include payments to prior servicers, then you'd be able to include some of the loans purchased in the last 12 months. Then 12 for 12 or better would be more than 70% of our entire portfolio. One of the things that's important to note, that loans that are 12 for 12 and weighted average LTV in the low 90s with a coupon of about 4%, now trade to a sub 4% yield. So, we basically buy loans that made six payments or seven payments and we pay 10 to 15 points less for those loans that then when they make 12 payments are worth approximately par. In some cases more than par. So, it's really one, the data analysis we do to figure out what loans we ought to be buying. And two, having the servicer being so close to us and be able to drive the strategies that cause the servicer to be able to enable borrowers to continue to reperform based on the loan characteristics that we have. So, with all this reperformance, it obviously increases cash flow pretty dramatically and as we just talked about, it increases kind of the implied NAV pretty dramatically. The other thing that significant outperformance does is it lowers our asset based financing costs over time. We're currently working on a couple of securitizations that will further lower asset based cost of funds, probably to the tune of about 50 to 75 or 80 basis points on about $300 million plus of loans. Keep in mind that a 50-basis point reduction in asset base cost of funds overall increases ROE by about 300 basis points based on the Q3 numbers. And on our metrics page, it's a little more identifiable. And lastly, another on the NAV front, we are looking at subject to REIT tax rules, selling some of our 12 for 12s or better paid history loans and likely to the extent that we do that, that would be a first quarter 2018 transaction. If we jump to page 10, financial metrics, this is a new table that we put in here. A number of our shareholders have expressed interest in us making this in a more formatted table rather than looking through the Q in multiple places. So, we decided to put it in here. One of the things you'll see is the average debt cost slightly higher and that's really the additional convertible notes we issued in August. Our leverage was down slightly, but average debt cost was slightly higher just because of the convertible notes. We would expect that number to decline over time, particularly as we do new securitizations. We're looking at calling a couple of our securitizations and refinancing them at materially lower rates. That would potentially cause us to amortize some of the issuance costs in deals we call, but it would be a fraction of the amount of money we would save in the refinancing. If you look at our non-interest operating expenses to average assets, it's a pretty stable number. It's basically 0.5% a quarter. That includes management fees and all other non-interest expenses. It's pretty stable number. The one thing I will say is that number actually declines with a little bit more leverage rather than increases, so you actually see as the assets grow, it becomes even more efficient. You compare that kind of on the community bank front, and I don't think you'll see too many loan balance sheets like ours where expenses are all inclusive basically 2% a year. The ROAE is also a pretty stable number. If you take a look, it increased a little bit over Q2, even with the impairment and then to the extent you see a lower cost of funds, that 12.7% number is actually closer to 16% or 15.5% than 12.7% to the extent that we can reduce our cost of funds through additional asset based funding refinances. Subsequent events, it's a pretty exciting time. We have a lot going on. We've already closed purchases of about $45 million of UPB this quarter. Very low prices to collateral value, a couple million of it is small balance commercial. We have another $10 million pending. We also have about $175 million that we are buying in a 50/50 partnership in a bond structure with Blackrock. So, you heard earlier that we closed our third round joint venture with DoubleLine in late September and then in early December we expect to close $175 million initial kickoff joint venture with Blackrock as well and that will be structured in a levered bond structure that we will each own 50% of. Dividend of $0.30 a share to be payable December 1 to stockholders of record November 17. And then the last thing I want to talk about in subsequent events, which is newer than this presentation, which isn't very old, is that subject to documentation and other requirements, Great Ajax Corp. has reached a preliminary agreement with the owners of Gregory Funding, our servicer, in which Great Ajax Corp. will acquire a 5% interest as well as it will acquire warrants to purchase another 15%. All of this would be new capital and Gregory Funding would not be buying out existing holders. And it would further align interests and also given the demand we see for Gregory services in our joint ventures, it would enable Great Ajax to have a lot of optionality in terms of the value of Gregory funding over time. Next two pages are balance sheets and income statements. And with that, I'm happy to answer any questions you might have.

We will now begin the question-and-answer session. [Operator Instructions] The first question comes from Jessica Levi-Ribner with B. Riley-FBR. Please go ahead.

Hi. Thanks for taking my question.

Sure. Jess, how are you?

Good. Thanks. How are you?

Good. Good.

So, just to go right back to your news that you're acquiring a 5% interest in Gregory, how can we think about that from an earnings standpoint and from a valuation standpoint, both for you and for Gregory?

Sure. And the outside Directors, I'll give them a lot of credit because they spent a long time thinking about how to structure it and initially hired Houlihan Lokey to do a valuation to kind of give them some guidance, but spent a lot of time going through numbers. One of the reasons why they decided to do only 5% was the same reason they decided that they wanted lots of warrants rather than only some warrants. The original discussion was buy a 9.9% interest and get no warrants. Then the more time they spent looking at and the more they saw demand for Gregory from people like Blackrock and DoubleLine, they decided that the optionality was really what mattered and that the best thing to do was buy a little less now in return for getting more optionality later. So, as a result, the 5% of $35 million, is a fair price. It's a little lower than what I will call the 50th percentile or right around the 50th percentile of the Houlihan valuation. But the upside from the warrants is really what drives it. From an earnings perspective, I don't think it will be hugely material in the next quarter or two, but as these JVs grow, it can become pretty material pretty quickly. But I wouldn't expect it -- realistically, the target is to get it closed by year end. There's obviously a lot of documents that have to get done. So, it wouldn't really be effective in any mathematical way to the income statement until the first quarter. And then my guess is that it wouldn't be material really until the second or third quarter. But when I think of from an NAV perspective, the valuation upside, that over time, if Gregory does what we all think it can do, I think the outside Directors will be glad to have those warrants.

And did you say 5% of $35 million?

So, it's based on the $35 million valuation of Gregory.

Okay.

And they will buy a 5% interest using that valuation and it will all be new capital, it won't be buying out a holder.

Okay.

And then they will have warrants. The prelim warrant numbers are $38 million valuation -- 5% at $38 million, 5% at $42 million, 5% at $50 million.

$42 million and then it jumps to $50 million?

Right. And that's the prelim handshake. Amongst the Gregory holders.

Okay. Thanks so much for that.

No problem.

Two other just slight questions on the expense line. The increased real estate operating expense, that was from that one NPL that you mentioned on the call?

Yes. So, about $1.1 million of impairment. One property was about 30% of that number and frustrating would be an understatement.

I hear you. But that should normalize or hopefully normalize on a go-forward?

Yes. REO is about as peaked as it's ever going to be from our NPL portfolios. So, yes, we'll see some REO over time, but the lion's share of the REO that's going to come from our late 2014 and early 2015 purchases, the lion's share has already been acquired and it is just kind of the remnants that are going to come last. Now, the good thing about the REO that comes last is the more the borrower fights, it's because the borrower is fighting because they like their house, not because they don't like their house. And as a result, the extreme tail of an NPL portfolio generally turns out to be okay or better properties and kind of the first two-thirds or first 70% tend to be the lesser properties.

Okay. And then one last thing, in the subsequent event area in the press release, you note that the majority of the costs associated with the acquisitions already done in the fourth quarter were accrued in the third quarter. Do you know how much that was about?

Off the top of my head I don't, but I have Mary here with me and she can probably look up that number. But it's probably about $600 per loan that closed plus $600 a loan that we kicked out. So, if you assume we have about 90% pull through, kind of a way to make a guesstimate would be take the number of loans and multiply it by 110% of $600.

Yes, we have a couple hundred accrued on the balance sheet right now. It could go up or down just depending on how close we are on the accrual. What we typically do, we say what work has been done, what did we pay, that's what's unpaid. But some of it will filter over into the fourth quarter, so the work has yet to be performed and then we have to do an additional accrual.

Right. So, if we agreed in September to buy something in October and we ordered the title policies and we ordered the compliance checks and all that in September, that would be a Q3 -- that would be a September expense even if we acquired the loan in October.

Okay, great. Thanks so much.

The next question comes from Scott Valentin with Compass Point. Please go ahead.

Thanks operator. Good afternoon. Thanks for taking my question.

My pleasure.

Just with regard -- you guys had mentioned last quarter potential sale of some 12 for 12 or 24 for 24s in the home loan market. Is there any progress there? Any thoughts on that going forward?

Yes, we've had specific discussions with a couple of buyers. Some of which depend on what loans we decide to put into our securitization structures when we select the final populations. And then some of the loans that we think would be better off in a sale structure. So, while I don't want to say we're definitely going to sell loans, I think it's a realistic possibility that would be in the first quarter. But we need to come up with a final population. The other thing is there's a lot of REIT tax rules as to how much you can sell over time in any one year and in any rolling three-year average period. And we want to make sure that we always have at least 50% of that available to us because you never know when you see an opportunity that is so mouthwatering that you want to sell something and able to go buy it. So, we always want to leave a little bit of buffer room in the REIT tax rule calculation. So, if we were to sell loans, my guess would be it would be less than $100 million in terms of total sale of UPB, but probably north of $50 million.

Okay, that helps. Thank you. And then I guess segue into the second question, last quarter you talked about achieving a rated transaction and that would lower the cost of funds even further. I guess that's still on track for the fourth quarter?

Yes. Still on track. We're working on exact timing. It could be early December, it could be early January. It won't be late December. We're also at the same time working on an unrated securitization and the order to some extent depends on some of the call rules of some of our 2015 B securitizations.

Okay. And so in addition to what's in the portfolio today, you would call some existing securitizations to the--

We would likely either call or buy some loans out from some existing securitizations. We're spending a lot of time with lawyers and tax people figuring that out a good part of everyday right now.

Okay. And then just final question on the dividend, I think you went through last quarter, had paid out $0.58, had earned $0.77 taxable income. I think again in this quarter you had $0.89 now of taxable income, paid out $0.88, so you're basically tracking exactly dividend to taxable income. Is that the way to think about it going forward? I know you wanted to keep the dividend kind of a base dividend, very manageable kind of a steady state upward march and then do specials every once in a while when it seems appropriate. Is that still the case, still the outlook?

Yes. The goal is steady state upward march in a predictable way. Unfortunately, performing loan taxable income is pretty predictable, but non-performing loan, because of the tax rules, is much harder to predict. Non-performing loans, as you can see from our portfolio, we've dramatically reduced as a percentage of the portfolio, but because of the absolute dollars of it from late 2014 and early 2015, it still has a material effect in any one quarter on taxable income. And the tax rules for non-performing loans were clearly not designed with REITs in mind.

Okay. All right. Appreciate it. Thank you.

The next question comes from Kevin Barker with Piper Jaffray. Please go ahead.

Hey Larry, thanks for taking my questions.

Hey Kevin.

Hey, how are you doing? Just had a quick question on the reasons behind the acquisition of Gregory, or at least the 5% from Gregory and the 15% warrants and the impetus for that occurring now.

So, if you're an outside Director and you know that based on the servicing contract that you have with Gregory Funding, that Gregory can't service for any third-party unless Great Ajax's outside directors approve it, and unless Great Ajax -- or unless Great Ajax is making an investment in the same portfolio. So, the outside Directors to some extent control the ability for Gregory to grow. So, if you saw from some large institutional investors some demand for Gregory's servicing products, you as an outside director would say, wow, I might be able to buy something really cheap and then give it permission.

Got it. And then given that this is occurring, would you conceive any other consolidation of the corporate structure as it stands right now? Whether it be Gregory or other parts of the Ajax infrastructure and how the corporate structure is set up right now?

Sure. I'd love for everything to be one and have it be completely internal. There is some regulatory restrictions and some ownership restrictions that made us set this up differently at the start. I think that our 20% ownership in the manager mitigates some of that, but the operating risk is still at the manager, but we get 20% of the ups basically with a zero basis. Having an interest in Gregory also helps get closer to that, especially with the optionality and the ability to have a say over its growth. I don't see anything near, near, near-term in the corporate structure that would change materially. But over time, it very well could, based on situations that we see. The one thing I will say is, for our outside Directors, it was absolutely, absolutely, absolutely important they were unwilling to make this investment in Gregory if it was going to be buying shares from the existing holder. They were only willing to do it if it was new capital.

Okay. All right. That's all I had. Thank you.

The next question comes from Robert Dodd with Raymond James. Please go ahead.

Hi. A couple. On the issue that you explained the delta between GAAP and taxable this quarter and actually selling the REO. If -- given obviously you're at peak now, but you'd expect to continue to be selling those properties. Is that elevated delta between GAAP and taxable going to continue from that source in the near-term?

I wouldn't expect it to be as extreme as it was in this quarter. This quarter saw an awful lot of REO timing of closings. So -- and some late Q3 stuff where the cash came in later. Title companies sometimes -- I know you find this hard to believe, but sometimes title companies, when something closes on a Friday, don't send the money until Monday and it could be a different month. But I would expect that this quarter was probably a little more extreme than would be the norm. I would see some volatility from REO sales -- the other place you see it is when you have nonperforming loans that you do loan mods on, the mod is considered for tax purposes a debt for debt exchange. So, that creates taxable income. So, the more non-performing loans that reperform, that creates taxable income. And the more non-performing loans that get foreclosed on, creates taxable income. But the more REO you sell after you foreclosed on it, it creates taxable loss. So--

Okay, that's really helpful actually. On kind [Indiscernible] looking into next year and maybe in Q1, selling some of your 12 for 12s, I mean would the intent be, to your point, there's rules about how much in a rolling period, et cetera, et cetera. Would the plan be to sell it in kind of a lump or spread it out? If you smear it out over a longer period, would that be easier to comply with the REIT rules, i.e., is it going to become a recurring quarterly thing or just kind of a one-time lump?

I think it's less likely to be a recurring quarterly thing. I think it will be more of a lump and it will likely coincide with the quick reuse of the proceeds.

All right. Okay, fair enough.

So, think of it as, if I can sell for $100 and buy for $81 and then have it be worth $100 12 months later, I want to do that.

Yes, that's a good trade. Last one, on the taxable question of that, does that, does the sale of a 12 for 12 reperforming, does that create a taxable gain or a taxable loss or anything like that?

So, it's likely to create a taxable gain with one caveat. That to the extent the 12 for 12s are loans that have already been modified, you've already had the taxable gain on that.

Okay, [Indiscernible]?

It's the full employment act for tax accounting.

So, beyond that, I mean on just the longer, obviously you've got a lot going on it looks like for Q4 and some sales in Q1 already. Are you seeing anything beyond that in the market that makes you optimistic or concerned about pricing or anything like that as you head into next year? I mean as you say, the pricing on NPLs has been elevated for a while. You've been keen on that pricing for a long time. Is there anything else you see in the market in terms of pricing and things like that that you think might be a little longer tailed?

Yes, I mean we've clearly seen banks, especially larger banks, being much bigger sellers than they have been in the past. And I think a lot of that has to do with home price appreciation has enabled them to be able to recapture reserves rather than take reserves when they sell. And it gets classified assets away and into someone else's hands and they get to recapture reserves. We've definitely seen funds, especially funds that were large NPL buyers, change their strategy dramatically from trying to liquidate as fast as possible to trying to modify as fast as possible, get paying and then sell. Depending on what they need in those sales, we found a number of funds who have been selling to us for years who have certain requirements like they need a mark or they need a certain amount of proceeds by a certain date. So, an auction structure doesn't work for them. So, we've actually seen more loans than ever before. It's a little bit lumpier than it used to be in the sense that while in Q2 we had a lot of transactions, in Q3 we had eight transactions, in Q4 we've already had a number of transactions and we expect a number more. But I think the size will continue to creep up of our individual transactions. But on the flip side, we're closing out a pool of one loan later this week. So, we're also seeing more activity on the small balance commercial side, which we also like. Although that will take a little while to ramp up and we have a lot of institutional investors who would like to co-invest with those, co-invest in those with us. So, the loan market is pretty strong. There clearly is a much better demand for things for loans that they believe they can go to the rating agencies and get immediately rated, almost like a rating agency enhancement level arbitrage game versus loans that they have to wait and get to perform until they get to that point, which is what our specialty really is. So, we haven't seen much price pressure. Maybe a point in our space. But in the clean pay 12 of 12 NBA current space, it's a sub 4% market now.

Okay, got it. Very helpful. Thank you. Thanks a lot guys.

Sure.

[Operator Instructions] A follow-up question from Scott Valentin with Compass Point. Please go ahead.

Thanks for taking my follow-up. I just want to confirm maybe the timing, I think you mentioned, I forget what page it was, it was the page that kind of showed performance metrics for the quarter, but you mentioned -- it was page 10. You mentioned with the lower cost of funds, you thought the ROAE, ex-REO impairment losses and lower cost of funds could get to 15% to 16%. I mean--

Yes, think of it as -- it's based on -- if you take our asset level finance, just our asset-based finance, not our overall finance, but just the securitization debt and the repo financing, asset level financing. If you were to reduce that by 50 basis points in costs and we would expect that over the course of late 2017 and 2018 that you'll see our securitization altogether financing reduced by about 50 to 70 basis points, our repo financing is obviously subject to where LIBOR goes. But our securitization financing, clearly where our 2015 and 2016 transactions were versus where our 2017 transaction in May and where it looks like these next two securitizations that we're working on will get done, will be significantly less expensive, perhaps as much as 75 basis points less expensive.

Okay. So, it would be fair to say, again, to your point about LIBOR, assuming it follows the current curve, you would be 15%, 16% ROE by say Q2 2018, ex-the REO impairments and things like that?

Yes. Assuming we get these two securitizations done and probably one more sometime in 2018.

Thanks. Very helpful. Thank you.

Sure.

This concludes our question-and-answer session. I would like to turn the conference back over to Lawrence Mendelsohn for any closing remarks.

Thank you everybody for being at our third quarter 2017 conference call. Mary and Russell and I thank you for joining and thank you for your questions and we're around and happy to discuss any additional questions you might have. And with that, I hope everybody has a good Tuesday night.

The conference is now concluded. Thank you for attending today's presentation. You may now disconnect.