Great Ajax Corp (AJX) 2019 Q1 法說會逐字稿

Good afternoon, and welcome to the Great Ajax First Quarter 2019 Financial Results Conference Call.

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

I would now like to turn the conference over to Lawrence Mendelsohn, Chief Executive Officer.

Please go ahead.

Thank you, operator.

Thank you, everybody, for joining us on our first quarter investor call.

To start off, I'd like everybody to look at Page 2 where we have our safe harbor disclosure and forward-looking statements.

And with that, we'll get into the presentation.

I apologize a little bit for my voice.

I'm fighting off a little of bit cold that seems to be going around Portland right now.

Overall, we had a very, very successful net asset value and intrinsic value-building quarter in pretty much most facets of our business and investment side.

We bought loans at good prices and very good prices with collateral values.

We added approximately $400 million in co-investment joint ventures with accredited institutional partners.

We closed 2 joint venture securitizations with good prearranged executions both in cost of funds and advanced rates.

And the joint ventures, you have to keep in mind, have significantly more effect than just the income generation.

Our servicer and our manager each get fees from those joint ventures, and we own a significant percentage of our servicer and our manager and their value increases materially as these joint ventures aggregate.

Just kind of a business overview from off top.

We're very, very, very focused on expanding our investor -- our seller base, our loan seller base.

Over 90% of our transactions continue to be in privately negotiated purchases.

We've made 273 transactions since 2014.

We had 12 just in the first quarter of 2019.

Our loan sourcing network is very, very important to our ability to acquire the types of loans we want at the prices we want.

Our sellers are typically banks, originators and funds who also invest in loans.

We use our manager's proprietary analytics to price each pool asset by asset by asset.

We frequently get the ability to go loan by loan with sellers and determine what subset we want to purchase.

We analyze a large amount of data to do this in order to determine target loan characteristics and to develop pattern recognition algorithms for both pricing and to drive the servicing of loans.

Third parties as well as some of our JV partners rely on our manager's analytics as a service.

We own 20% of the equity of our manager at a 0 cost.

As a result, it doesn't show on our balance sheet.

We adjust individual loan prices to accumulate clusters.

We have -- we purchased in the first quarter some very small pools.

We also, in joint ventures, purchased very large pools.

We aggregate much lower cost bases than our competitors, and our affiliated servicers then service these loans asset by asset, borrower by borrower.

And the servicer's performance really has created brand value for the servicer and tremendous amount of NAV, which we'll look at later when we see the migration patterns of our loans.

And the servicer's performance has led institutional investors to us for loan purchase joint ventures and third-party servicing.

In Q1, we increased our JVs by approximately $400 million and between Q4 of 2018 and Q1 by $1 billion.

This materially increased the servicer's value.

We own 20% of the servicer including warrants.

We invested in the servicer about a year ago, in the first quarter of 2018.

And since then, our servicer's portfolio has increased by 50%.

We use moderate non-mark-to-market leverage.

Average leverage for Q4 was 3.3x and asset base leverage was 3.0x, pretty much unchanged from Q4 -- Q1 2019 was pretty much unchanged from Q4.

We'll see more of this when we look at some of the comparison tables later on in the presentation.

Now a little more quarter focus, Q1 2019.

We purchased about $28 million, $29 million of loans separate from JVs at a purchase price of about 61% of property value.

Almost $18 million of those were small-balance commercial loans generally in urban areas, particularly in Southern California, Houston and Dallas, Texas.

We formed -- we also formed joint ventures that acquired almost $400 million in mortgage loans with underlying collateral value of $670 million.

We retained $64 million of the varying classes in the related securities that we do in structuring, and we closed the quarter with $183 million of investments in these joint ventures.

These JVs we closed in -- that we closed in first quarter of 2019, all $400 million closed in the last 2 business days of March.

So those were the 28th and 29th of March.

As a result, they were on our balance sheet for an average of 2.5 days.

So we received very little income from them in Q1, but we're glad we closed them and the income from them will show up in Q2.

We acquired one multifamily property for $2.3 million.

That's in Atlanta, Georgia.

We're very excited about that one.

Interest income, $29.5 million, net of $3 million (sic) [$0.3 million] in servicing fee expense.

I talked about that a little bit in Q4, but I wanted to bring it up one more time.

Interest income from our portion of the JVs shows up in income from securities, not loans.

Also, servicing fees for securities are paid out of the securities waterfall.

So our interest income from securities is net, not gross of servicing fees unlike loans.

As a result, as our JVs grow, it causes interest income to be lower by the amount of the servicing fees from our joint ventures.

In Q1 2019, this difference is about 80 basis points on average invested amount in securities.

Another place you can see the JV effect on interest income is if you look at the servicing fees line item on Page 13 of this presentation, you'll see that servicing fees actually declined for the quarter about $50,000, and that's because they don't show up in the securities.

They show up as effectively less interest income in the securities.

So that's a little bit of a distortion, but it's created because of the specific JV structure of securities that we do with our accredited institutional partners.

Basic earnings of $0.39; net income, $0.39.

There are few items I want to mention that were small offsets to net income.

We've discussed before REO impairment was about $500,000 or about $0.025.

We mentioned on previous quarterly calls REO impairment happens first under GAAP and the gains happened second.

Also, any foreclosure that is -- results in a third-party sale of the property rather than becoming an REO is accounted for as a loan payoff and not an REO sale and doesn't offset any REO impairment in our income statement.

In our portfolio, third-party sales happen approximately 50% of the time.

So 50% of the actual REOs that get sold don't get sold as REO.

They get sold through loan payoff effectively at a foreclosure sale.

Second, we had an increase in professional fees of approximately $280,000.

Approximately $250,000 of this was a onetime accrual for SOX testing as this is our first year of no longer being an emerging growth company under SOX rules.

So this charge for the SOX testing -- upfront SOX testing and the opinion is about $250,000, about $0.015 a share.

We had another $0.01 of other onetime charges related to specific loans in older, small remaining pools.

Since we don't mark loans to market, a small decrease in a few loans becomes a GAAP charge and isn't offset by an increase in market value of our larger loan portfolio.

We'll talk about the market value of our loan portfolio when we get to the migration table later on.

So with the $0.05 of noise, we walk back to approximately $0.44 a share.

With the -- including the $0.05 of noise, $0.39 a share.

Taxable income, $0.11.

This is an unusual number.

There's a couple of reasons for it why it was lower.

First, we had fewer foreclosures.

Foreclosures generate taxable income at the time a foreclosure becomes an REO, equal to the gross value of the REO, property minus tax basis in the loan.

We had more REO sales than new REO created through foreclosures.

REO sales typically generate tax losses since the foreclosure that creates the REO causes taxable income based on gross REO value before all expenses related to the REO.

The sale of the REO, you have expenses that typically you recognize when you sell the property and it generates a tax loss.

So if you have more sales than new REO, you generally have net tax loss.

Three, fewer prepayments in Q1 partially because of the timing here, partially government shutdown, partially some volatility in rates in late Q4 and early Q1.

However, we have started to see a pickup in prepayment in later March due to the February rate rally.

There's usually about a 45- to 60-day delay of the prepayment from rate rallies, and we're definitely seeing in both April and May so far that more loans are paying off than in Q1.

And number four, more loans are paying every single month.

We'll again see this in the migration chart.

But for paying loans, GAAP income is -- from our purchase price discount is determined by the expected life of the loan, while taxable income is determined based on installments to contractual maturity.

Since contractual maturity is typically much longer than expected life, for paying loans, taxable income is usually lower than GAAP income.

Prepayment accelerates taxable income, so there's an automatic catch up.

And loan sales would also accelerate taxable income, and that's something we'll talk about a little bit on this call as well.

We collected $63.2 million of cash.

This is approximately 17% annual rate based on our investments in loans and securities.

It shows how many loans are paying, but it also shows that prepayment is down a little because in Q4 2018, this number was about 18.5% versus 17%.

17% is still an enormous number, but it represents the difference in prepayment between Q4 and Q1.

Average cash for the quarter was $55 million.

Quarter-end cash was $41.5 million.

On Page 5, our portfolio overview, the composition is not much different.

The RPLs remain 97% of our portfolio and NPLs, 3% of our portfolio.

REO is principally held for sale.

It turns into cash over relatively short periods of time.

REO increased marginally from Q4 of '18 to Q1 of '19, but not from foreclosure or from fewer REO sales.

Instead, we purchased a commercial property.

As our property portfolio grows, we'll see an increase in total REO, but not REO held for sale.

We'll also see an increase in noninterest income.

On Page 6, you can see kind of the relative ratios of where we buy things.

We continue to buy lower LTV loans with overall RPL purchase price of approximately 62% of property value and approximately 86% of UPB.

The price to property value does not include home price appreciation, if any, since acquisition.

You can see that our price to property value has been pretty steady, all along the way.

We're very focused on always playing defense and always having kind of a base expectation of return, and we stick to our knitting in that regard.

Purchased NPLs had been declining in absolute dollars invested.

For our NPL portfolio, purchase price to property value is approximately [56%].

As you might imagine, higher LTV NPLs become REO sooner and lower LTV NPLs become REO later, if at all, as lower LTV NPLs are much more likely to become paying loans or just pay off.

You can see at 3/31/2019, at the end of the quarter, our purchase price to property value is about 55%, and that's been a pretty consistent number for a long time.

Again, that doesn't include any home price appreciation in the underlying properties since our acquisition time.

On Page 8, we have a couple of changes to our portfolio concentration.

California, however, continues to represent the largest segment of our loan portfolio.

It increased pretty significantly at the end of the year.

Right at the end of December, we bought a pool of loans in a joint venture structure that was 89% California, about $240 million.

Our share of it was 20%, and it's the largest segment of the portfolio both for residential RPLs and for small-balance commercial mortgage loans.

Our California assets are primarily Los Angeles, Orange and San Diego counties.

We're seeing consistent payment and performance patterns in these markets particularly in urban centers.

We also have seen consistent prepayment especially for certain borrower characteristic subsets in those markets.

The California percentage of our portfolio increased in Q4 of '18 and it also increased in Q1 of 2019.

You'll see in Dallas and Houston markets, they now have purple in them, not just target markets, but we now have property management.

We've added to our Houston and Dallas investments as well as our infrastructure in those cities.

One thing we are seeing in certain markets, particularly in New York, New Jersey, Connecticut and Illinois, is an effect of the new tax law.

We've seen a compacting of higher and middle range home prices were higher end have come down materially more than middle and upper middle, and we were seeing a much smaller difference between middle range home prices in those markets and higher end prices in those markets as well.

Portfolio migration.

This is one of the things that we are extremely, extremely focused on.

It's the real basis for NAV creation here, the way we buy the loans, the way we think about the loans, we want to buy the analytics we do and the way we drive the servicing of these loans based on analytics.

If you look at our portfolio right now, $1.25 billion of it is 12 of 12 consecutive payments or better, and that's an increase of $110 million versus December 31, 2018.

When we bought these loans, only $150 million of them were 12 for 12.

For the loans that we buy based on analytics and the way we -- our servicer services and manages these loans, our data suggests that once a loan becomes 7 of 7 consecutive payments, there is about a 93% chance for the loans that we buy that they will become 12 of 12 consecutive payments.

At acquisition, less than 10% of our loans were 12 of 12 or better.

Now the overwhelming majority of our loans are 12 of 12 or better.

The intrinsic value of loans has increased materially on average since acquisition.

Some of you may see some of the home loan markets where loans have traded once they're 12 of 12 clean pay, rating agency eligible and it's night and day versus our average purchase price in the mid-80s of these loans.

So it's a significant NAV creator.

Number two, in addition to increasing cash flow and NAV of our balance sheet, the significant outperformance of these loans also over time lowers asset base cost of funds and increases the senior securitized bond advance rates.

We've clearly seen this in our JV structures where we get 75% and 80% senior advance rate where market standard is more like 65% senior advance rate.

In first quarter of 2019, we did not issue any non-JV securitizations, but we are working on a couple of them currently.

Structured credit markets are materially better than they were in first quarter as well as in fourth quarter of last year.

And we will -- we expect that 1 or 2 securitizations that we do in the second quarter will materially lower cost of bonds on the related financing for those loans.

On Page 10, subsequent events.

We've already closed since March 31 about $14 million at a $88 price, 57.8% of property value.

We have another, about $30 million under contract, 3 of which are properties: 1 in Dallas, 1 in Baltimore, 1 in Raleigh, and about $22 million of which are re-performing loans.

You can see the purchase price on the RPL is 84% and the price/collateral value, 56%, which is over and over and over as our network of sellers and the analytics we do that it allows us to be able to do this.

We've had a pretty busy Q1 so far.

There's a lot of things that we'll be looking to offload some assets before June 30, and it's still pretty early in Q2.

But it's important to see the pattern of we just keep buying loans at approximately the same prices with the similar composition and collateral value percentage.

Our Board announced -- declared a dividend, $0.32, payable on May 31, 2019, to shareholders of record on the 17th.

You may think about that relative to $0.11 of taxable income from our Board's perspective given the underlying NAV of our assets.

They're pretty confident that the taxable income number versus the dividend will be -- the dividend will be maintained as is.

So the -- with that, we'll go to Page 11, financial metrics.

And this is -- there's 2 slides.

The first one is where I'll focus, which is excluding the consolidation of a couple of our joint ventures.

We have 2 of our joint ventures which we are still required to consolidate.

We own more than 20% of those joint ventures, which make it a little confusing, but there's a couple of topics I want to mention.

Average loan yield, you'll see, is 8.7% versus 8.5% in Q4.

Yield on loans is net of any impairment.

So in Q4, we had more impairment than in Q1, but we still had some impairment.

You'll see in Q3 we had almost no impairment, so the yield is higher.

So our yields on our loans have been relatively constant.

Impairment's the only deviation in any of the quarters for the most part.

If you look at average yield on debt securities, that 7.3%, remember that is net of the servicing fee of approximately 80 basis points.

Debt securities is how our interests in our JVs are presented for GAAP accounting.

As our JVs increase, which they have materially in Q4 and Q1, the GAAP reporting of that servicing, unlike loan interest income, distorts the average asset yield lower and the related ratios to that.

So the average total asset yield of 8.5% is also distorted a little bit lower because of the 7.3%, which is net of about 80 basis points of servicing relative to the cost.

If you look at our average debt cost, it's really the difference is 2 things.

One is rounding between a number just over 5.1 and a number just under 5.2.

Number two is we put on some 6-month repurchase agreements on a non-mark-to-market basis in the fourth quarter to deflect any movements in volatility that we saw in late fourth quarter and first quarter.

We put them on 6-month agreements.

6-month LIBOR was higher than it is now.

And when those roll, which they are in April and May, repo funding will come down by about 20, 25 basis points on those related repos.

If you look at our ending leverage and our average leverage, they're pretty consistent ending leverage.

It's the same as it was at the end of the year, 3.6x including convertible -- our convertible debt and our asset-backed debt.

The convertible debt's unchanged versus year-end.

The asset-backed debt is 3.2 versus 3.3.

With so many loans paying every single month, as we saw on the migration chart, we're comfortable with a little more asset base leverage, although in Q1, total leverage didn't increase and our asset base leverage was basically the same as well.

But we do have capacity and room and comfort given our portfolio performance of increasing asset base leverage a bit more.

With that, the last 2 pages are our quarter end financial statements, and I'm happy to take any questions anybody might have.

(Operator Instructions) The first question comes from Tim Hayes with B. Riley FBR.

This is actually Mike Smyth on for Tim.

So the first question I had, I was wondering if you could provide a little more color about the amount of RPLs purchased.

It's a little bit less than what was announced for 1Q on the prior earnings call.

So I was just wondering if this is due to some type of fallout or if the pipeline was just getting pushed out due to volatility?

Or if you can provide any other color here would be helpful.

Sure.

It was --- what we closed at the end of March was -- had some due diligence fallout based on seller documents that we thought would potentially make loans not enforceable in certain states.

And as a result, we choose not to purchase them because of that.

So really due diligence more than anything else.

Got you.

And how often does that happen?

Is that something that we should be kind of...

It depends -- different sellers, we have different pull-through rates.

If you look at, for example, December 31 in our 2018-F transaction, there was a prefunding account for about $100 million of nonperforming loans, and we pulled through $60 million of those in Q4 because the documentation from the seller was so poor that we decided that it wasn't economically intelligent to make the acquisition of 100% of the loans.

So that's -- having 40% fallout is an enormous amount.

Typical is probably 3% to 5% fallout.

Got you.

That's helpful.

And then another question, I was wondering if you could talk a little bit about your outlook for JV formation for the remainder of 2019.

So noticed the pace was just a little bit down relative to the fourth quarter.

So I was wondering if this is volatility related or just due to timing, or if you can provide some additional color here, that would also be helpful.

Sure.

Sure.

The demand for joint ventures has increased dramatically.

We have seen a little bit more, for larger pools, prices increasing.

And as you can tell from kind of what we do, we are very not just priced to UPB, but extremely priced to collateral value, and we want specific kinds of collateral.

So we have relationships with a bunch of big banks and funds who we have the ability to negotiate with, and one of the things we've seen is they all have huge pipelines of what they want to get rid of.

So it's really a function of when they're going to -- when they're sellers as opposed to when we're buyers.

Got you.

That's...

But we've been told by 3 banks that they, over the next 18 months, have combined somewhere between $25 billion and $40 billion that they're going to sell.

Got you.

That's helpful.

Of RPLs.

The -- we've also seen that the agencies are large sellers, and we also have a couple of aggregation structures.

We're actually in the documentation phase of a $300 million aggregation structure for small-balance commercial mortgages.

Got you.

Is there any other color you could provide there?

I would expect that we'd be in buying -- ready for buying probably sometime in mid-June.

And yes, it's us and an accredited institutional investor.

We'll be 25%.

They'll be 75%.

And we have a credit facility ready to go for when we get the documentation done for that structure.

Awesome.

Sounds good.

That's also great color.

And then just one last question.

I was wondering if you can provide an update on Gregory.

So you mentioned it's -- the value has increased since the investment.

But I was just wondering if you can provide any color compared to the fourth quarter and just kind of how revenues are trending compared to the fourth quarter as well.

Sure.

Revenues at Gregory continue to increase.

We've also expanded on the software development side at Gregory to build out technology even more especially in the tenant and property management side.

We expect additional joint ventures and third-party servicing to continue to increase over time.

We've been approached by a number of third parties actually seeking to make investments in Gregory, which kind of -- maybe we should've expected it but we didn't really.

So all the things we keep hearing about Gregory are all positive, and the results they're having on loans is remarkable.

Awesome.

That's good color.

Would you be able to provide a number on the change in value compared to the prior quarter?

It's like me telling you, "Oh, this is what I think Uber is worth", right?

I don't know other than -- Gregory is -- effectively, 20% of the economics are owned by Great Ajax and another 25% -- and then 2 institutional investors have about 20% or 25% each and then the original Aspen also has about 25%.

And I don't know how each party values their own investment on a mark-to-market basis.

But my own feeling is that the valuation of Gregory is probably at least 50% higher now than it was in first quarter of 2018.

And the question is different -- based on different strategies, we would consider different -- I think Gregory would consider different perhaps investment ideas from third parties.

But we've been approached by a number of -- Gregory has been approached by a number of them and it's been interesting.

The next question comes from Scott Valentin with Compass Point.

Just with regard to the margin, I guess it declined obviously from several quarters ago but seems to be stabilizing.

Is that a fair outlook going forward?

The market should run at 3.4% level, 3.5% level?

Yes, I think that over the next quarter or 2, cost of funds will come down a little because 6-month repo in Q4 was a lot more expensive than where repo would be or replacing it with securitized funding would be.

So I would think that cost of funds would come down a little bit in Q2 as well as Q3, and I think that yield on loans is pretty stable.

Okay.

Okay.

That helps.

And then on the SOX expense, is that a onetime event?

Or is that something you'll incur the next several quarters?

Yes, it's -- well, it's part -- that's a portion of a larger amount, but the first year is more than successive years.

Okay.

So that will be the first -- so that $250,000, that represents the first year's expense?

Yes, first year's -- no, the $250,000...

It's the incremental cost.

Obviously, as Great Ajax grows and adds more joint ventures and more joint venture partners, our audit fees are increasing because we're becoming more complex and there's more to audit.

But not -- but -- so if Great Ajax were to grow unbelievably in the next 12 months, it would be $250,000 more expensive.

So maybe a small part of that will be ongoing but not the entire amount for sure.

The CFO of Great Ajax failed to properly accrue throughout the year, so you can blame me for that one.

But it was more funding an emerging growth company.

Fair enough.

And then just, Larry, you talked about the embedded value in the company, and I think a lot of investors have acknowledged that the servicer, the investment manager, there is value there.

But how does the Board intend to unlock that value?

I mean I know you guys can speak to it, you can put it on paper, but I think investors are trying to figure out when does that -- when is it possible for that value to get unlocked?

Sure.

Well, a couple of things.

One, just in the value created by the servicer from the migration of the loan portfolio.

Stranger things have happened than us maybe selling some loans because if you continually buy loans in the 80s and when they become 12 of 12 [at] par, at some point, you ought to be a little bit of a seller at par and continue to reinvest in the 80s.

So I think that's one of the things our Board would say.

Now in REIT, rules restrict how much you can sell on any given year and things like that.

So it's a -- there are limitations and restrictions, but that does make economic sense.

From a straight manager servicer perspective, I think the REIT board, when they made the investment, they made it as part of allowing them to go service for -- the servicer to go service for other people and servicing these joint venture structures.

That really kicked off the joint ventures.

And since that, just kind of the Board permitting Gregory to do that as part of the initial investment in January of 2018, those joint ventures are probably $1.5 billion or $1.6 billion since then, actually maybe a little more than that.

And from commitments from joint venture partners, if we can find all the assets, that could probably triple over the next year or 2. And really it's going to be more of a function of us finding the right assets and continuing to service them in the way we service them.

And then number three, we've had a number of people reach out to us purely just as third-party servicers because of the way they've seen our securitizations in loans and our securitizations perform.

So kind of that combo and, for lack of a better term, kind of brand getting around, we've had a number of firms reach out about making investments in Gregory to allow it to get bigger and to also potentially acquire other servicers because of the ability to create efficiency and stronger analytics.

So the Board and the management of the servicer are going to be busy over the next 18 to 24 months.

Okay.

All right.

And then one follow-up question just on other income.

I guess is there -- that's grown dramatically from a year ago.

It's been roughly flat linked quarter but still at a pretty high level.

What's flowing through there?

Is that service-related income or -- I think $1.1 million -- I think it's $1.1 million last quarter, but it's up from like $400,000, $500,000 last year, prior 2 quarters?

Sure.

Some of it is rent coming from properties.

Some of it is late fees.

As our performing loan portfolio grows, we're -- one thing that's different about Gregory's contract and Gregory's mentality is typically you see the servicer gets late fees.

And one of the things we've never understood is why should the servicer be better off because if the borrower pays the 17th of the month than the 1st of the month, you want everybody to pay the 1st of the month.

So as a result, all late fees go to Great Ajax.

They don't go to the servicer.

So as the portfolio expands, late fees obviously expands from paying loans -- the more loans that pay every month -- that pay between the 17th and the 31st, Great Ajax makes extra other income.

So it's the rental income as well as late fee income that comes through the other income item.

(Operator Instructions) The next question comes from Kevin Barker with Piper Jaffray.

Mary, on the -- in the interest income and some of the stuff up there, how much of the interest income was related to accretion versus actual cash?

It seemed like [there's accretion in certain periods], but I know you explained some of it earlier.

But can you give a little more color around that?

50-50 these days.

Yes, it's...

Cash is probably greater.

Cash is actually greater than accretion.

Yes.

Okay.

How much was accretion this quarter offhand?

Mary?

[In the Q] tomorrow.

Mary is here and looking through the key draft, but it will actually be detailed in the Q when it gets filed tomorrow.

Yes.

You'll see in the statement of cash flows, you'll see the noncash portion.

And then on the accretion tables, you'll see the total.

But accretion -- the accretion is a much smaller percentage than cash.

Right.

So it would seem like prepayment fees were relatively high this quarter, right?

No.

They were actually pretty slow in Q1 relative to what you -- Q4 was slow, Q1 was a little slower and some of that was the volatility in Q4.

Some of it seasonality.

We see a lot less prepayment in December and January because of holidays.

We also, with the government shutdown, it was a little bit dysfunctional mortgage market.

But one thing we have seen is since about late March, early April, prepayments have picked up pretty materially particularly in certain states like California.

So is that on NPL portfolios that have been performing for well over 12 for 12 of added capacity to refi...

What we're seeing -- we're seeing it in 2 places.

We're seeing it in paying RPLs that are 12 for 12.

The other place we're seeing it is we're seeing empty nesters selling.

And in different states with different absolute dollars of equity, the thresholds are different.

So -- but these certain specifications we've identified that we can forecast percentages of empty nesters selling now based on backtesting.

We've had a bunch of loans that had been [modi-ed] 4 years ago by a previous owner that were 2.5%, 3%, 3.25% interest rates and we couldn't figure out why they were prepaying.

So we started backtesting them and we found them to be -- have commonality of empty nester, and then we started backtesting those characteristics of the empty nester to come up with a pattern of which loans would prepay that had arbitrarily low interest rates.

Well, that will be a good time for the housing market.

All right.

All of my other questions have been answered.

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 very much for joining our first quarter financial results conference call.

We appreciate your time and listening in the questions.

And if you have any questions, feel free to give us a call or reach out to us.

We're always happy to talk about Great Ajax.

And with that, I hope everyone has a good evening.

The conference has now concluded.

Thank you for attending today's presentation.

You may now disconnect.