Great Ajax Corp (AJX) 2017 Q2 法說會逐字稿

Good afternoon, and welcome to the Great Ajax Corp.

Second Quarter 2017 Financial Results Conference Call.

(Operator Instructions) Please note, this event is being recorded.

I would now like to turn the conference over to Larry Mendelsohn, Chairman and CEO.

Please go ahead, sir.

Thank you very much.

Thank you, everybody, for joining Great Ajax's second quarter 2017 conference call.

We appreciate everybody on the line.

Before we get started, I want to point out the safe harbor disclosure for forward-looking statements on Page 2 of the presentation.

And with that, we can start talking about our business and the quarter we just concluded.

Kind of at the top level, in general, a pretty good quarter in terms of the underlying economics of our business, NAV creation, portfolio growth and lowering funding costs.

But we also did have, as you see, some GAAP noise from calling our 2015-A securitization a year early.

That was $0.01 to $0.015 share, and REO accounting requirements similar to fourth quarter 2016 of about $600,000.

There's also been continuing positive developments in loan markets as a whole and some extremely positive developments in the securitization markets over all.

With that, on Page 3, I'll give you a quick business overview of where we are.

We continue to have over 90% of our acquisitions since inception being privately negotiated and Q2 was more of the same, except 19 transactions, a big change over first quarter of this year.

Our sourcing network is really important in our ability to acquire the types of loans we want in the markets we want them and the prices we pay.

When we get to the highlights page, I'll talk a little bit more about the loans that we bought, and you'll be able to see really how they fit our grand master plan.

The manager's proprietary analytics, large amounts of data.

We spent a lot of time analyzing where we want loans, what markets, what MSAs and as well as what target characteristics of those loans we want and in terms of forecasting performance.

Affiliated servicer Gregory Funding looks at loan-by-loan and asset-by-asset in terms of bringing it out.

Just as an aside, I know we mentioned a few times that the outside directors have hired Houlihan Lokey to do an evaluation on the servicer about potentially making a small investment in the servicer.

Houlihan Lokey delivered to the outside directors their report, and the outside directors have formed a committee to evaluate it and we expect to have some prelim feedback from that committee before the end of summer.

We use -- we like moderate leverage.

We don't like extreme leverage.

We really like non-mark-to-market leverage, which is how we structure most of our repurchase agreements and all of our securitizations.

This quarter, we issued some convertible debt, in late April of 2017.

So we now separate what I'll call asset-level leverage, which is securitization or repo financing versus corporate leverage, which also adds in the convertible bond.

So just at the asset level, it's about 2.63x.

So leverage did not increase that materially at the asset level and corporate leverage of about 2.91x, including the $87.5 million of convertible bonds.

That being said, we use convertible bond proceeds to acquire a lot of loans in the second quarter, so we have a fair number of unlevered loans as well as cash on hand.

We've done 9 securitizations since our inception.

We did another one, 2017-A in the second quarter.

We like nonrecourse fixed rate funding.

We have a great group of bond investors.

I can't say enough about how helpful and positive they are.

Many of them either co-invest directly in loans with us or we're in the process of putting that together with some.

The 2017-A securitization, which I just mentioned, included some new loans as well as all the loans in our 2015-A securitization that we called early.

We had increased the leverage on those loans and decreased the funding cost versus 2015-A by more than 1% per year on all those loans.

On Page 4, I'll talk more just about the quarter.

Probably the biggest highlight is we -- purchase price of $210 million on almost $250 million re-performing loans.

Even more remarkable than that is the underlying collateral value of $357 million on those loans.

So when you think about 19 transactions, the purchase price of 84% of UPB, but even more remarkable, the purchase price of 58.7% on the underlying property value.

And we'll talk a little bit more about that as we look at our -- more portfolio details in the next few pages.

We issued $87.5 million of convertible notes, 7.25% coupon, convertible at $15.37 in the quarter.

We also issued a 2017-A securitization.

It was $140 million of senior bonds at 3.47%.

The net increase in debt of that $140 million was only $98 million since we called our 2015-A deal and we paid down the related debt for those loans as well as some related debt for some of the other loans in the securitization.

Probably the most interesting thing about the securitization, aside from it's a 4 percentage point cheaper than the 2015-A, was of the $140 million of senior bonds, $50 million actually funded a pre-funding account for loans not yet acquired.

And it was an account valued for about 30 days of future acquisitions, and we funded about $49-point-something million of the $50 million for loans bought after the securitization but directly funded into the securitization.

Pre-funding accounts are something that really haven't been seen in a very long time, and we're really happy that our bondholders have enough confidence in us that they're willing to create bonds to secure -- secured by loans not yet purchased and gave us 30-plus days to add to the underlying mix.

From a net income basis and an interest income basis, interest income of $21.7 million; net interest income, about $12.4 million.

That was a rise partly due to the loans we bought.

On the flip side, the early call of our 2015-A deal, we had the remaining $218,000 of deferred issuance costs.

We called the deal 1 year earlier than I think is mandatory, and we accelerated the amortization at $218,000.

Instead of taking it over the following 12 months, we took it all up front.

We had $600,000 REO impairment.

I'll talk more about that as we get through the presentation with some detail on Page 10, but it's very similar to our December impairment.

These are different REO.

They came in over the first 6 months, and you'll see that we have significant built-in gains in the other REO.

And then close to $1 million of just plain timing difference issuing a convert of $87 million in late April and closing loans in late May and June.

So the amount of interest you earn off the loans is less than the convertible, especially since the loans weren't yet levered up to increase the number of loans relative to convert in a short period of time.

So that's really a timing difference.

We' see that change over Q3 and Q4 pretty materially, particularly as we continue to acquire loans.

If we jump to Page 5, you can really get a feel for our portfolio.

Last quarter we were about 93% performing loans.

Now 95% of our portfolio was re-performing loans, 5% nonperforming loans.

If you look at from a property value perspective, we have significant excess property value relative to our loan portfolio.

And in REO front, we have $42 million of market value REO.

That's all REO, both REO held-for-sale, which we'll see on Page 10; but also some REO held for rental, which is typically multi-unit small-balance properties that we have on our balance sheet.

On Page 6 we have a little more detail about our re-performing loan portfolio.

For RPLs, we continue to be very focused on buying low LTV loans.

We've started that in February of 2015 and have continued.

And the overall purchase price of our RPL portfolio is 62.8% of the initial underlying collateral value and 80.8% of the UPB.

Just as an aside, at December 31, our overall purchase price was 64.8% and now it's down to 62.8% because of purchases we made this year.

So we continue to lower our purchase price relative to the underlying property value.

It's sort of playing offense and defense at the same time, and given the strength in housing prices, it's something that we think makes a lot of sense.

And you'll see, when we talk about taxable income, that this low LTV approach is having material effect on increasing total cash flow from the loans and also accelerating prepayments.

On the nonperforming side, it continues to shrink by about 20% quarter-over-quarter on a carrying basis.

Our purchase price to collateral is 55% on nonperforming loans and to UPB, about 62%.

Purchased NPLs continue to decline as a percentage of our portfolio and in absolute dollars, and we think we have a pretty low cost in them.

That being said, we still don't believe that nonperforming loans going forward are as good an investment opportunity as the re-performing loan market is.

From a portfolio perspective, the map hasn't changed, although there's probably one little change that we could add and that's in Houston.

While we are close to adding Houston back as a residential market, we have added it in as a small-balance commercial loan and property market for our portfolio.

And we are working on some loan acquisitions on some small-balance commercial and some property as well in the Houston market.

California continues to be our biggest market.

Nearly 30% of our overall portfolio is in California, and 75% to 80% of that is in Santa Barbara, south of San Diego, so Southern California.

We still like the dynamics of that market.

We like the attractiveness of the time line in case the loans stop performing and the stability of property values in that market as well as the job market.

On Page 9, I'll talk a little bit about portfolio migration.

We introduced this last quarter to give you a feel as another way of thinking about NAV, but also kind of how we buy loans and what they turn out to be and what they turn out to be worth.

So we focus much more on buying loans that have made somewhere between 4 and 7 and 4 and 8 payments.

And this is a chart that shows what those loans have done after we've owned them.

So this chart shows payments to us.

It does not show payments made to previous servicers.

So if we bought a loan that was 7 of 7, it wouldn't show up as 12 for 12 here until it made 12 payments to us.

So it would have actually made 19 payments in that case.

So this really just shows payments to us.

Because of that, anything that's 12 for 12, we would have had to own for 12 months already.

So if it wasn't purchased prior to July of 2016, it couldn't be 12 of 12.

One of the things this shows is that about $530 million of our portfolio was now 12 for 12 or better to us.

That's up $55 million from quarter end March 2017, so the end of last quarter.

It's up $55 million quarter-over-quarter.

If you were to add-in payments to the previous servicer as well, that 12 for 12 number would be over 70% of our portfolio, which is actually pretty remarkable performance.

Now even then, it still wouldn't take into account anything really bought this year because there couldn't have been more than a few payments to us because of the time it takes to servicing transfer.

So for example, the $250 million of UPB we bought this quarter couldn't possibly be in this table at even -- at 12 for 12 and barely at even 7 for 7. Only a small portion would be at the time of acquisition.

So that's #1.

If we think about what this means for our portfolio of 12 for 12, large pools of 12 for 12 now trade at extremely high prices.

We actually saw a $900 million portfolio traded last week at about 106 of the accruing balance accreted to a sub-4% yield.

And we saw another pool of about $400 million trade at a approximately 4% yield.

When you think about our portfolio of loans that we bought somewhere between 4 of 4 and 7 of 7, we didn't buy those anywhere close to a 3% and 3.25% or 4% yield.

So we have a material built-in gain when these things become 12 of 12.

And in some cases, if you look at our purchase price of approximately 81-or-so on UPB, it could be as much as 10 or 15 points in some cases.

The -- when we buy 6 of 6s and 7 of 7s and become 12 of 12s.

And when they become 24 of 24s, they trade at a very high price.

So part of just the value creation is loans paying.

Part of loans paying is our ability to analyze data target-specific markets and target-specific loan characteristics that we think will add to payment.

Okay.

But in addition to increasing just the NAV of the underlying loan portfolio, these payments actually create a significant amount of cash flow.

And when you think about cash flow, we collected $44 million off of our portfolio in this quarter.

When you think about what $44 million means on an annualized basis, that's $170-plus million or almost 17% of the entire carrying cost of our portfolio, which is a significant amount of cash flow from a loan portfolio with a 4.5% coupon.

We've definitely seen that have several effects for us.

One, aside from increasing NAV, increasing taxable income, increasing cash flow, it also has helped us lower our cost of funds.

We saw just in this quarter we were able to, on the re-securitized and newly levered loans, reduce our cost of funds by 1% a year.

We're working with the rating agencies on a rated deal structure, with a target some time in Q4 that we think would further lower our cost of funds on another $200 million-or-so of underlying loans, perhaps as much as 1.25% to 1.5% and also provide more leverage than we currently get in our unlevered -- or our unrated securitizations or our repurchase agreement facilities.

And lastly, when we talk about where 12 of 12s and 24 of 24s sell in the open market, subject to REIT rules, we may very well sell some of our 12 of 12s.

Our board's given us permission, subject to the REIT rules and the annual and 3-year averages, to go out and explore selling a small percentage of our portfolio that comply with the REIT rules and redeploy the proceeds in more 4 of 4 to 7 of 7 loans.

On Page 10, as promised, I want to talk about the real estate owned that's held-for-sale.

This page does not include real estate that is out for rental, that is currently being rented.

This is just our real estate owned held-for-sale.

You can see that on a net liquidation basis, we expect to have about $34 million of net proceeds.

That's not a gross number.

That's a net number versus our carrying value of about $29.7 million.

So even after the $600,000 impairment, we expect that we have a built-in gain of about $4.4 million.

That's actually a little bit higher than our expected built-in gain of about $4 million from December.

So in our overall REO portfolio that's for sale, we believe we have material built-in gains.

That being said, GAAP doesn't allow us to use built-in gains to offset built-in losses.

So we take the built-in losses as we determine them rather than as we sell properties.

GAAP requires us to take our gains when we sell those underlying properties.

As you might imagine, our experience over the years with, especially, with nonperforming loans is that the REO that you take back first rather than last is the tail.

So the tail is better REO, and what comes first as worse REO, it tends to be higher LTV loans with less engaged borrowers, either low or negative amounts of equity.

That's what tends to become REO first.

They tend to be in worst condition, and they are generally the lower relative market value properties in whatever MSA they're in as REO goes in our portfolio.

For the very best REO, with the most equity in a foreclosure, it usually sells to a third party after foreclosure sale.

That is not in REO sale.

For our accounting, that's actually a loan payoff because we never take the REO.

So that gets buried in pool accounting as a loan payoff.

So the REO that you'll see here is either the tail, which happen first, or it's the second best REO rather than the first best REO that you usually see come on the nonperforming loans.

That being said, GAAP doesn't allow you to mark to market the good, only the bad, unless you're willing to mark to market 100% of the assets on your balance sheet, and we think that, that would be volatile in any given time and would be less informative about our overall company performance.

On Page 11, similar to looking at portfolio migration and performance as a way of thinking about NAV, we also say let's let a true third party, the bond market, the structured credit market decide what our portfolio's worth.

And we took our last securitization that we did in May of this year, May 25 at close, and you'll see that we're actually able to increase our subordinated bonds by another 2%, up to 12% total versus 10% total.

We still hold all our subordinated bonds.

We haven't sold any subordinated bonds yet.

And we kept the senior the same, but at 3.5% or 3.47% instead of 100 basis points higher all-in yield.

So this is all updated, our entire portfolio, as if we had securitized them in our last 2 securitizations and what the structured credit market forecasts our portfolio was worth.

It has -- kind of an interesting sidebar as to what it means to return on equity, if you think of return on equity, if we were to sell our subordinate bonds and just keep the residual, the equity trust certificate which is equal to 23% of the UPB of the portfolio, the equity basis would be approximately 3.5% of the portfolio.

So we would own 23% of the UPB with 95% RPLs that are outperforming expectations on a cash flow basis, and we'd own them for 3.5 points.

So on a kind of implied leverage basis, it's effectively the equivalent of having 96.5% leverage if we were to sell our subordinate bonds, our B1s and B2s and still own 23% UPB.

Also, given where we've taken kind of an average of where we think the equity trust certificate is worth based on some third parties, we've had people offer us higher than 40%, but we use that number as a -- in the presentations historically.

We didn't want to put in a higher number and just add to the NAV.

But if you look at it, it's almost $3 a share, about $2.89 implied NAV increase over our $15.49 book, which implies about $18.38 of implied NAV.

From structured credit, again, this is not our opinion.

This is another way to look at it just like portfolio migration where we're actually looking to third-parties to give us an understanding of what the implied NAV of our portfolio is.

Subsequent to quarter end, as usual, a lot going on.

We have about $30 million of RPL acquisitions that will close in the next few weeks, 88 loans in 2 transactions.

Again, you'll see purchase price about 80% of UPB; price to collateral value, about 63%.

You can see that the loans are underlying around 75 or 80 LTV.

So again, focused on lower LTV loans and low purchase price at the underlying collateral value.

Those closed.

We also have -- I'm sorry, those were July closings.

And then in August, we have another $3 million scheduled to close.

There's actually for the summer a lot of loans that are out there that we're seeing.

We're a little surprised by it.

It's not usually what we see.

Summer tends to be a slower quarter on the acquisition front, but it seems to have a little more energy than we traditionally see.

On the small-balance commercial front, we've already closed $1.7 million.

We're looking at 3 more small-balance commercial loans of another $3 million, $3.5 million, and we would expect to see more funded this quarter as well.

As part of the small-balance commercial piece, we're in the documentation phase with a large institutional fixed income investor on a JV to grow the small-balance commercial platform.

And with any luck in lawyer cooperation, that will move forward pretty quickly.

So let's talk about dividend.

We talked about cash flow, we talked about taxable income being $0.39 and in Q1, taxable income was $0.38.

Combined, that's $0.77.

Our board decided to raise the dividend to $0.30; record date, August 15; pay date, August 30.

$0.30 plus $0.28 last quarter is $0.58.

As a REIT, you know we're obligated to pay under REIT rules 90% minimum of taxable income.

90% of $0.77 is more than $0.58.

Our board wants to be predictable, doesn't like to make large jumps in dividends.

Taxable income is a little harder to forecast than GAAP income is because of level yield accounting.

It's not how tax works.

So as a result, our expectation, not a guarantee, but our expectation is that dividend will have to go up as taxable income continues to increase from cash flow and that absent some material change in taxable income, we would expect the trend in the dividend to continue.

On Page 13 and 14, our income statement and balance sheets.

This is the first quarter that basic and diluted are slightly different because we have a convertible bond, again, that we issued in late April of 2017.

And with that, I'm happy to open up to anybody with any questions and happy to talk about what we're doing.

(Operator Instructions) And your first question will be from Jessica Levi-Ribner of FBR.

This is Tim for Jessica, and thanks for the color on the dividend policy.

It sounds like -- I pretty much kind of know the answer on this one already.

But -- so it's seemingly that this quarterly dividend is likely to keep going up so long as there's this big discrepancy between taxable income and the dividend just based on REIT rules.

But is that something that you and the board would kind of rather see that quarterly dividend go up and trend up rather than do a special kind of true-up at year-end?

I think our bias is to have the quarterly dividend thread -- kind of thread its way up, work its way up, rather than have a special dividend.

I think there's a realistic chance that there could be some special dividend either way this year, but our expectation based on our most recent board meeting a few weeks ago or a week ago is that the directors expect that the dividend will continue to move its way up rather than have just a giant jump.

Okay.

And you mentioned a couple initiatives in terms of REO sales and potential 12 for 12 loan sales.

And so between kind of the capital freed up from those asset sales and the cash you have on hand and what's left from the convert, just what's your -- how much dry powder do you have available today?

And what does that mean for your -- kind of what's your capacity around acquiring RPLs and investing in SBC loans right now?

Sure.

If you just take kind of our unlevered loans that we could always put debt on and our ability to get some leverage from a rate of securitization and cash on hand, our RPL powder by itself is probably another $150 million.

So if you think about just the convert itself, that probably gave us the ability to buy somewhere between $300 million to $400 million.

Assuming we levered it in a securitization structure at 3x, we traditionally get about 3.5x to 3.9x in a senior bond, but if you just assume 3x, that gets us somewhere around $350 million-or-so.

So we have plenty of dry powder.

We have ETM in place.

We've never used it.

We have no plan on using it.

And I think the convert was done at -- one, it was done for a specific reason because we had $200-plus million of loans under agreement to be purchased, and we wouldn't think about raising capital absent having another $200 million or $300-plus million to buy.

Okay.

That's really helpful.

And then one more, just around the JV you had mentioned.

Is there any preliminary talk as to kind of size or contribution to that?

And if you could just talk about the SBC market a little bit, the dynamics you're seeing versus kind of larger balance stuff and why you like the space.

Sure.

We -- from an SBC perspective, our space is really what I'd call $500,000 to $5 million, but even more so, probably $500,000 to about $3 million.

It's predominantly inner city.

It's predominantly has a residential component, may not be straight multi-family.

It may be ground floor retail with apartments above, some kind of mixed-use.

We tend to do what I'd call moderate transitional lending: someone's acquiring a building, they need 60% loan and they need another $500,000 over 6 months to renovate some of the units or to get -- use money to get retail tenants, things like that.

We think it's a pretty attractive market.

It works well in many of the markets that we're already concentrated in, in our portfolio: L.A., Phoenix, Houston, Miami, Atlanta, Washington D.C. and Bronx, Brooklyn, Queens and Portland, a few other cities.

From a pure size perspective, we think it's a pretty significant market and one that we can ramp up in those cities.

And we want to have concentration because just like everything else, we'd like to have our own people on the ground, and having specific targeted markets and concentration helps, that makes it -- turn out better.

From a JV perspective, each party is looking to contribute $75 million of equity, and we have 2 investment banks that we're pretty far along in discussions that are willing to provide 2:1 leverage against that.

I think that to fully fund, if each party did $50 million, that would be $100 million levered 2:1.

So, call it, $300 million.

That's -- in full throttle, that's probably a 2-year investment cycle, not a 1-year investment cycle; or maybe an 18-month investment cycle, but it's not a 12-month investment cycle.

And if we do it as $150 million and $450 million, that probably makes it a 2.5-year investment cycle, absent buying any pools of small-balance commercial.

That would just be straight kind of semi-origination where we're providing the financing either at the time of the loan or shortly thereafter in the next few weeks.

There are community banks that were under a lot of pressure to actually exit that market from the regulators.

So there is opportunity to buy pools of those loans also, and we're actively out talking to banks that we deal with and that we get refer to by some of our institutional investors for discussions about buying some of their small-balance commercial.

The other thing we've had discussions with a number of banks is kind of the built-in leverage concept of the A and B structure where we actually own a loan and sell a senior participation to a small bank and at a lower rate.

And effectively, it's built-in financing for us and for the bank, it's a much better capital treatment, senior debt.

So that's another avenue we're exploring with a number of banks in these markets as well.

The next question will be from Steve Delaney of JMP Securities.

To echo Tim's comment, thank you for the dividend hike.

We had not projected that until the fourth quarter, so you're making us look good here, which we like.

So listen, I heard something new and maybe because I wasn't listening close enough last quarter's call, but I'm intrigued by the conversations the board is having about the possibility of purchasing an interest in your dedicated servicer, Gregory.

Could you just clarify that?

I know -- it sounds like it's nothing you can talk about in specifics.

But where did that initiate?

Is it the board of Great Ajax that is making that request?

I'm just curious how that all came about, but it sounds like it would be a very shareholder-friendly move with respect to Ajax shareholders.

I think that is true.

The outside directors of Great Ajax started asking about the possibility of investing Gregory probably about a year ago.

And one of the things that made them really interested, aside from the fact that they think it's a good alignment of interest because Great Ajax -- or Gregory is captive, they saw that there were a number of institutional structured bond investors who wanted to buy loans with us and have Gregory be the servicer.

And as a result of that, they thought that as Gregory is captive, that if there was going to be a time to invest in Gregory, it would be before you allow Gregory to go do that rather than after, number one, because you could put Gregory across the inflection point.

Number two, they look at it as kind of a strategic investment that gives Gregory the ability to invest more in its own infrastructure as well as cause Great Ajax kind of the Gregory's -- the interest that these institutional investors -- structured credit investors have in using Gregory as a servicer has enabled Great Ajax to see a lot of, one, loans that these institutional investors may have been seeing that we were not; and two, it's made the institutional investors be willing to be more aggressive in our securitizations.

For example, you don't see too many $50 million pre-funding accounts in securitizations that are 6x or 7x oversubscribed to buy loans that we haven't yet purchased, right?

That's something -- that was [in 2008]...

That was something they did back in the CLO days, right, pre-crisis?

That's exactly right.

And so you don't see that, and our structured credit investors have enough confidence in us buying loans and in Gregory servicing them that they want more bonds and are willing to do a prefunding account because they know what we will buy and they know how Gregory will service.

So it gives us execution benefits in our bonds.

It gets us to see loans that we may not see that the large institutions do see.

It gets them to want to do those JV, partners with us in order to get Gregory to then service those loans for them.

So the outside directors, they're all very -- one of the things we did when we set up Great Ajax, so we wanted really sophisticated, business savvy directors.

And these directors kind of look at it and say if there's a time to do it, let's do it, and they hired Houlihan Lokey probably about 5 or 6 months ago, I think in January, to do evaluation that was delivered to them a couple of weeks ago.

At the board meeting last week, they formed a committee of outside directors to work directly with Houlihan and that evaluation to put together what they think they're willing to do.

And they preliminary told us -- preliminarily told us that they would come back approximately late August.

Okay.

Great.

That's wonderful color and background.

I really appreciate that.

And just from where I sit as an analyst, I think that would be -- we were aware of the institutional accounts that you have that are going on and assumed that Gregory was doing work for those other parties.

So it does raise the question of who's benefiting there or possibly could the attention be diverted.

So we get it, yes.

And then if you could on the securitization market, and we're seeing obviously credit spreads come in, in a lot of markets, but you wouldn't think the RPL market would be necessarily the first one to tighten so much.

But just in the very general sense, Larry, so if you look at where you are today versus maybe let's go back to the 2015 deal you just called, you're paying more for loans today on a price to UPB.

But my question is, if you look at the big picture, would you say that the terms within the securitization market have largely offset the higher cost base that you have in your loans and therefore a lower yield?

Are you getting close to the same sort of targeted levered ROEs in today's securitization market that you are able to get, say, 1 to 2 years ago when loans appeared to be cheaper on an unlevered basis?

Sure.

So if you look at kind of apples-to-apples of where we're buying loans right now versus where we bought them 2 years ago, and when I say apples-to-apples, 2 years ago, we were buying a little bit higher LTV loans versus now we're buying less -- lower LTV loans.

So some of the purchase price increase is really driven by the safety of the loans we're buying as opposed to changes in loan prices for, say, 4 of 4s or 7 of 7 loans.

The biggest changes in loan prices are really in the clean pace that are 12 of 12 or better or close to 12 of 12, 11s of 11s, something like that.

The securitization market itself however, especially on senior bonds, have tightened enormously.

For example, if you look at rated transactions, AAA securitization -- AAA bonds now, you see trading at 85 over treasuries on a 3-year -- 3.5-year basis.

So you're talking about bond yields of like 2 65 on AAAs.

And you're seeing leverage on BBBs, in some cases, all the way up to 80% of UPB and the BBB maybe in the mid-3s.

So you're seeing all-in is people getting 75%, 80% of UPB.

Not of cost of UPB in the low 3s, plus expenses.

So -- and we look at it -- our securitization market has gone from basically 65 at kind of 4 90 all-in and we're on unrated basis 2 65 at 3 90 all-in.

So we're saving -- and that 65% UPB, if you think of our cost of, say, 80 of UPB, I think our all-in cost is 80.8% on UPB.

If you think of our cost on re-performers of 80.8%, what that says is that we're getting about 4.5x leverage at 3.9% fixed all-in, including expenses, amortization of issuance costs, everything, and the 3.47% senior bond.

In a rated structure, we think we can reduce that by another 0.5 points and actually increase leverage from 65 to -- in the low 70s, and that's of UPB.

So that would be, call it, 6 to 1 leverage in the low 3s plus expenses or high 2s plus expenses, so maybe at 3.5 all-in at 6x leverage.

So the securitization right now, I'm...

And you're thinking -- you're talking about -- did I hear you say earlier, you're talking about possibly a rated deal the end of this year?

Yes, in Q4 is kind of what we're targeting.

Unfortunately, we don't get to decide when we do it.

Working with the rating agencies is really -- this is our first transaction.

We also don't want the first one to be the biggest one because the first one has the worst execution, not the -- of all these deals we do.

But our expectation is it gets significantly better than our unrated deals.

It won't have any step-up requirements.

It will be permanent financing.

It'll be match-funded.

And we think it lowers our all-in cost an additional 0.5% on top of the 1% we already saved on the last deal.

So it's -- the securitization market, the structured credit markets are really tight.

Part of it is, there is no supply.

I mean, if you think about kind of nonagency mortgages, it's a market that in 2008 was $2.8 trillion and today is $400 billion or something or 400 -- but shrinking by $10 billion or $12 billion a month, net.

So it's -- there's very little new supply and a lot of demand for yield.

And the one thing I will say is all the investors that we do within our structured bonds are pretty smart, thorough, diligence people and they get the difference that Gregory makes and that our data analytics and buying strategies make versus kind of just buying billion-dollar pools and doing things.

The next question will come from Robert Dodd of Raymond James.

Most our questions have been answered, but on the potential for the 12 for 12 to start selling, and you mentioned that the board's given approval, et cetera, would it be fair to say that once that -- and obviously as you pointed out, as we [talk] how long if we held things exactly, would it be fair to say that once that starts, that would become an ongoing process?

Or is the talk looking at a pool selling that and then stepping back away from that market again for a while?

Or is it just going to be...

I think that they see it as part of a overall strategy where you securitize some, you finance some, you sell some as opposed to necessarily targeting a specific amount of selling.

The REIT rules, holding time period requirements and you have the 3-year average total sales of, I think, 10% with no more than 20% in any 1 year.

And you never want to be up to your limit because in case you see something that makes your mouth water, you need to be able to sell something in case you have to, to redeploy.

So we always want to have some cushion.

That being said, if the board does like where the 12 of 12 and the 24 of 24 market is transacting right now versus where we buy loans and that they have assets to explore somewhere -- a sale of somewhere between 6% and maybe 7.5% of -- so to give us at least 3% or 4% built-in room under the REIT rules of our portfolio to explore it, so we'll start working on that.

Probably in August or early September, we'll actually start putting together prospective pool.

We want to have -- our preliminary information back from the rating agencies before we were to decide what that prospective pool would be.

Okay.

Got it.

And then one I'll ask on seasonality of acquisitions, you gave some color that Q3 looks like it's more active than normal for summer period.

But I mean, stepping back for a second, I mean, you look at Q3 last year, it was huge, right, but that was quite an anomaly, right?

The Q3 the year before was not as huge.

That's right.

There was an anomaly of some -- of a specific -- 2 specific sellers looking for timeliness of liquidity.

Right.

So if I can put you on the spot, not for specific numbers, but in a hypothetical year, how much -- if you were going to buy $100 million a year, roughly what would you expect -- it's not going to be 25, 25, 25, 25, right?

I mean, where would you expect the kind of the relative scale and activity to kind of shake out on average, not any given year, but...

Sure.

Third quarter and first quarter are the 2 slowest.

And some time -- and when we have surprises of volume, it's usually in the third quarter, not the first quarter.

Usually, a positive surprise in the first quarter is because something didn't get done year-end because the seller couldn't do something and then it gets into the first quarter.

We usually see acquisitions pick up kind of March through July.

So while we're a little bit in August, September tends to be a little quieter and then we see October, November, December busy.

January, usually the first 2 weeks are busy.

After that, it's slow until the 1st of March.

Got it.

Very helpful.

That's one of the tricky parts of the business, right?

And then...

Some of it has to do with timing of bank capital ratios.

Some of it has to do with sellers being around versus being on vacation.

Some of it has to do with what sellers are outraising new funds.

Some of it has to do with what sellers are trying to earn and carry by a certain date.

Some of it has to do with sellers that are trying to get a mark.

Some of it has to do with banks' regulatory capital ratios.

For example, we get a call from a bank and they need execution by a very odd date.

It means that their regulators are coming in for an exam 1 week or 2 after that odd date.

Right.

Right.

Yes, I got it.

And I'd ask you one on the dividend and again, thank you for the color.

And you mentioned the possibility of a special because obviously, the rate of tax was growing and the rate of the dividend is.

There may be a special.

Would you -- and this is maybe more detail than you want to give, but would you expect if a special were to occur, it would occur in calendar '17 or would it be in '18, right, because obviously you've got some timing -- you have flexibility on shareholders (inaudible) benefit?

Right.

Our expectation based on preliminary discussions with the tax folks is it likely would be December rather than January.

The next question will be from Kevin Barker of Piper Jaffray.

I just had a -- could you please go back through how you get to the $0.58 number?

You quickly went through it at the end of the prepared remarks.

But can you just go back to the...

The taxable income versus the dividend?

So dividend in Q1 was $0.28, dividend in Q2 is going to be $0.30, so that's $0.58.

Taxable income in Q1 was $0.38 and Q2 was $0.39, so that's $0.77.

Requirement so far this year would be 90% of $0.77, or $0.70.

And so far, we'll have paid out $0.58.

Got you.

So obviously, there is a big difference between what you have there and what the future would look like, right?

And then...

And if cash flow then can continue at that rate, that number gets -- that difference gets bigger rather than smaller, correct.

Okay.

And I appreciate the comments earlier which hones in on the timing.

And then in regards to your NAV, I notice in the disclosure it's been a couple quarters where we've seen a decline in the NAV.

Can you help us understand what the puts and takes and why the NAV is declining and then your outlook for it?

Sure.

The number one reason is our focus on much lower LTV loans.

So as a result, we're paying a little more as a percentage of UPB less on property value, but more on UPB.

And since the securitization structures are based on percentages of UPB, that actually gives us a little less leverage rather than a little more leverage from senior bonds.

So as a result, the built-in NAV is a little bit lower because of the kinds of loans that we'll be buying.

That being said, it's a little misleading because the kinds of loans we're buying are cash flowing so much more than the loans we used to buy and -- which makes them, from a total cash flow perspective, significantly more cash flow generating and would make the senior debt go away faster.

So -- which would increase the PV of the resid.

That's why I said based on the new loans we buy, we've had people who would value that resid materially higher than $0.40, but we kept that $0.40 so that it's consistent.

I would argue that, that $0.40 on the resid value is probably closer to $0.60 or $0.65, maybe even $0.70 in today's loan pricing market.

So it's entirely a reflection of the change in strategy towards cleaner credits and price-related difference, yes, and obviously that will change the discounted cash flows or the potential cash flows that could be generated from those differences.

Right.

So what would happen is we haven't changed that $0.40 value on the resid, but every $0.10 on the resid value is $1.5 of NAV.

And that concludes the question-and-answer session.

I would like to hand the conference back over to Larry Mendelsohn for his closing comments.

Thank you, everybody, for joining our second quarter 2017 earnings conference call.

We'll be around for most of the rest of the summer.

And if you have more questions, feel free to contact us.

We're always happy to talk about our business.

Appreciate it again.

Thanks for the support over the years, and we look forward to talking to you in the future.

Thank you.

Ladies and gentlemen, the conference has concluded.

Thank you for attending today's presentation.

You may now disconnect your lines.