AGNC Investment Corp (AGNCO) 2012 Q2 法說會逐字稿

Good morning and welcome to the American Capital Agency Q2 2012 shareholder call.

All participants will be in listen-only mode.

(Operator Instructions)

After today's presentation, there will be an opportunity to ask questions.

(Operator Instructions)

Please note, this event is being recorded.

I would now like to turn the conference over to Katie Wisecarver in Investor Relations.

Please go ahead.

Thank you, Amy.

And thank you for joining American Capital Agency's second-quarter 2012 earnings call.

Before we begin, I'd like to review the Safe Harbor statement.

This conference call and corresponding slide presentation contains statements that, to the extent they are not recitations of historical facts, constitute forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995.

All such forward-looking statements are intended to be subject to the Safe Harbor protections provided by the Reform Act.

Actual outcomes and results could differ materially from those forecast due to the impact of many factors beyond the control of AGNC.

All forward-looking statements included in this presentation are made only as of the date of this presentation and are subject to change without notice.

Certain factors that could cause actual results to differ materially from those contained in the forward-looking statements are included in the risk factors section of AGNC's periodic reports filed with the Securities and Exchange Commission.

Copies are available on the SEC's web site at www.sec.gov.

We disclaim any obligation to update our forward-looking statements unless required by law.

An archive of this presentation will be available on our web site.

And a telephone recording can be accessed through August 20, by dialing 877-344-7529.

And the conference ID number is 10015859.

To view the Q2 slide preparation, turn to our web site, agnc.com, and click on the Q2 2012 earnings presentation link in the upper right corner.

Select the webcast option for both slides and audio, or click on the link in the conference call section to view the streaming slide presentation during the call.

If you have any trouble with the webcast during the presentation, please hit F5 to refresh.

Participants on today's call include Malon Wilkus, Chairman and Chief Executive Officer; Sam Flax, Director, Executive Vice President and Secretary; Gary Kain, President and Chief Investment Officer; Chris Kuehl, Senior Vice President of Mortgage Investments; Peter Federico, Senior Vice President and Chief Risk Officer; Bernie Bell, Vice President and Controller; and Jason Campbell, Senior Vice President and Head of Asset and Liability Management.

With that, I'll turn the call over to Gary Kain.

Thanks, Katie.

Thanks to all of you for joining us today.

As we will discuss on today's call, our portfolio remains very well-positioned for the current market environment.

The current landscape is one where interest rates are at record lows.

The yield curve is considerably flatter.

And prepayments on most parts of the mortgage market are poised to exceed the high scene over the last few years.

Mortgage spreads relative to swaps and treasuries are tighter than where they've been in the recent past.

But valuations are generally warranted, given the interest rate and prepayment outlook.

Furthermore, the odds of additional quantitative easing from the Fed have increased since last quarter when we reviewed the three scenarios that we were focused on.

If we do get a QE3 that incorporates substantial purchases of agency MBS, we would expect lower coupon mortgage valuations to richen substantially.

And we would also anticipate the prepayment picture becoming even more challenging, with only a few places to hide.

As Chris will go over when he reviews the portfolio, we believe we have positioned our portfolio in the areas that will perform well, even if we do end up in this kind of environment.

But as we discussed on our last call, I really want to be clear.

We are not betting on a QE3.

This is evidenced by our hedge ratios and our somewhat lower leverage.

But we just cannot afford to be caught offsides if economic activity doesn't pick up and the Fed does decide to act.

It would be irresponsible for us to be unprepared for this scenario.

With respect to the second-quarter financial results, we feel good about AGNC's performance.

We continue to create significant value for our shareholders by producing total economic mark-to-market returns of around 22%.

While at the same time reducing our risk posture via improved asset selection, increased hedge ratios, lower leverage.

And by continuing to increase the average term of our repo funding.

With that, let's review the highlights of the quarter on slide 4. Our net comprehensive income, which incorporates unrealized gains and losses on both assets and hedges, was $1.58 per share.

Which was comprised of a GAAP net income loss of $0.88 per share and a gain of $2.46 per share in other comprehensive income.

Net spread income, which is what many refer to as core income, came in at $0.94 per share, essentially reversing the increase we saw last quarter.

When we exclude an $0.11 per share estimate for premium amortization catch-up cost, we get an adjusted net spread income of $1.05 per share.

The decline from last quarter's $1.30 in adjusted net spread income was largely a function of faster prepayment projections, reduced leverage, increased hedge ratios, higher refill costs and changes in asset composition.

With respect to the higher prepayment estimates, which were a big factor, it is important to note that our actual prepayments were unchanged quarter over quarter.

And actually declined on a monthly basis during the quarter.

As such, if we had used our actual prepayment speeds, our net spread income would have been higher.

Taxable net income, which is not impacted by the changes in the projected prepayment speeds, or by unrealized gains and losses on derivatives, continued to be solid at $1.62 per share.

Our undistributed taxable income increased considerably again to almost a $0.5 billion, or $1.61 per share as of June 30.

Even after adjusting this number for the share issuance in July, our undistributable taxable income equates to $1.44 per share post equity raise.

Book value increased modestly to $29.41 per share.

And, additionally, the performance of lower coupon HARP and lower loan balance mortgages has been very strong so far during Q3.

So we are seeing a material tail wind for book value at this point in the quarter.

Now, before I turn to slide 5, I want to briefly discuss our current thinking with respect to the undistributed taxable income in light of our continued book value growth.

When we made the decision to lower the dividend to $1.25 earlier this year, we said we believed the dividend should be sustainable for a reasonable period of time, given current market conditions and our best estimate of future financial performance.

Given the strong book value performance and the large amount of UTI, we continue to believe our dividend is sustainable at its current level over at least the near term.

This, despite the recent spread compression.

Now ultimately, however, future dividends will be a function of actual financial performance and are therefore uncertain.

Turning to slide 5. Our mortgage portfolio decreased to $78 billion from $81 billion by quarter end, as we reduced leverage from 8.4 times where it was at the end of Q1.

Leverage during the quarter averaged 7.5 times, while quarter end leverage came in at 7.6 times.

Our average CPR remained unchanged at 10% during the quarter.

But was only 8% in the most recent prepayment release in July.

Our quarter-end net interest spread decreased to 162 basis points, driven in large part by the combination of faster prepayment projections, increased hedge ratios and higher repo costs.

Lastly, we raised a small amount of accretive common equity via our ATM program.

And did our first preferred stock offering during the second quarter.

We also raised around $1.2 billion during July in a follow-on offering.

Now let's turn to slide 6 and look at what happened in the bond market during the second quarter.

As you can see, we had a significant rally in longer-term rates during the quarter, with the 10-year dropping 56 basis points.

The yield curve also flattened materially as the two-year treasury rallied only 3 basis points.

In price terms, the five-year treasury rallied around 1.5 points, while the 10-year treasury increased in price close to 5 points.

As shown on the bottom left, the prices of lower coupon mortgages also increased materially, but in response to the declines in interest rates.

While the price changes were more limited as you move to higher coupons.

The same trend can be seen on the top right with respect to 15-year mortgages.

Now if we turn to slide 7, we can see how AGNC's book value has performed over the past 3.5 years in some different interest rate environments.

The first thing that's important to point out is that 10-year rates have been pretty volatile over this period.

And despite that, AGNC's book value has increased in 13 of the past 14 quarters.

In aggregate, book value has increased 71% since the beginning of 2009, with book value growing from $17.20 to $29.41 per share.

This is, of course, in addition to the $18.85 of dividends we have paid during the same period.

Interestingly, as you can see in the bar chart on the bottom, our two weakest book value quarters were Q3 2010 and Q3 2011, both of which were characterized by substantial declines in interest rates.

Why?

There are a number of factors including the timing of equity raises and accretion.

But a key driver is that AGNC hedges a significant portion of its interest rate risk.

And mortgages often under-perform when interest rates fall quickly.

In particular, the premiums or pay-ups on slower prepaying mortgages tend to lag behind changes in interest rates, as the market generally requires some time to adjust to a changing prepayment landscape.

So what are the key takeaways from this slide?

I think our multi-year track regard demonstrates it is possible for an actively-managed portfolio, with the right assets, coupled with appropriate hedges, to not only produce industry-leading returns over time, but to do so with greater consistency and less volatility.

So now I'll turn the call over to Chris to discuss the changes to AGNC's portfolio during the quarter.

Thank you, Gary.

On slide 8, we've updated a familiar graph, highlighting the prepayment differences between various types of 2011 30-year 4% pass-throughs.

You can see that pools backed by loans with lower loan balances, as well as loans refinanced through the HARP program, continue to perform extremely well relative to more generic TBA pass-throughs.

Which have meaningfully sped up despite being a relatively new production cohort.

As Gary reviewed with you on the market update slide, interest rates are again at all-time lows and therefore asset selection is more critical than ever.

Given today's low interest rate environment, we expect to see prepayment speeds on more generic TBA pass-throughs continue to diverge from HARP and lower loan balance securities.

Many Wall Street investment firms expect speeds on TBA 30-year 4%s to be well into the high 30s CPR to low 40s CPR over the next several months.

On the other hand, lower loan balance and HARP securities will likely increase by only a few CPR.

As we turn to the next slide, keep these speeds in mind.

On slide 9, we have two hypothetical yield tables that illustrate the importance of prepayment speeds and their impact on returns.

In the table on the top half of the page, we have generic 30-year 4%s, which are currently prepaying in the mid-20s CPR.

As we discussed on the prior slide, these securities will likely be paying closer to 40 CPR over the next few months.

Look what happens to your ROE as prepayment speeds increase.

At a 20 CPR, we have an asset yield of 2.25% in this example.

If we assume a cost of funds of 70 basis points, that leaves a net spread of 1.55%.

If we apply 7.5 times leverage, that generates a gross ROE of 13.88%.

Look what happens at 30 CPR.

The asset yield drops to 1.46%, which leaves us with a net spread of just 76 basis points.

And with 7.5 times leverage, your ROE drops to 7.16%.

On the table in the lower half of the page, we again have 30-year 4%s.

However, here we're calculating the yield at a price 2 points above the TBA price.

This is roughly where some higher LTV HARP securities were trading as of June 30.

You can clearly see that despite the strong price performance of these strategies over the last few months, there's still a lot of value, even at today's higher payouts.

At 7.5% CPR, which is roughly where these securities have been paying, we have a gross ROE of nearly 19%.

Even if speeds increase by 5 CPR, which is beyond what we would expect in the current environment, the returns are still above 15%.

And very compelling relative to more generic TBA 30-year 4%s.

Given today's low interest rates and high dollar prices, prepayment speeds are going to be a primary driver of performance with or without a third round of QE from the Fed.

Let's turn to slide 10 to review the composition of the investment portfolio.

During the quarter, we increased our holdings of lower coupon HARP and lower loan balance pass-throughs to nearly 70% of the total portfolio.

But equally as important, is that more than two-thirds of the portfolio's other non-HARP loan balance positions are in low coupons.

In the case of 15-year, these positions are mostly 2.5s.

And in the case of 30-year, mostly 3.5s.

You can see a more detailed breakdown of these positions by coupon in the appendix on slide 25, where we give you the percentage of HARP and lower loan balance by coupon.

Lastly, I'd like to highlight the fact that the portfolio's prepayment speeds remain well contained.

And actually fell during the quarter despite speeds on more generic securities increasing.

With that, I'll turn the call over to Peter.

Thanks, Chris.

Today I'll briefly review our funding and hedging activity.

I will begin with our financing summary on slide 11.

The cost of our repo funding increased 5 basis points during the quarter.

At quarter end, our average repo cost was 42 basis points, up from 37 basis points the prior quarter.

This increase was driven predominantly by a general increase in repo costs.

And to a lesser extent, by the longer average maturity of our repo funding.

During the quarter, we further reduced rollover risk by extending the average maturity of our repo book.

At quarter end, the average original days to maturity of our repo funding was 121 days, up from 104 days the prior quarter.

As you can see from the table, a significant portion of our repo funding now includes maturities in the one-, two-, and three-year sectors.

To put this maturity extension in context, just 12 months ago, our repo funding had an average maturity of 20 days.

Since that time, and consistent with the growth of the portfolio, we have termed out a significant portion of our funding.

Thus reducing rollover risk and developing greater access to longer-term repo funding.

A summary of our hedge positions is provided on slides 12 and 13.

Slide 12 shows the detail associated with our swap and swaptions portfolios.

Given the decline in interest rates experienced during the quarter, and the changing composition of our asset portfolio, we increased the size and duration of our pay fixed swap portfolio.

With swap rates reaching historical lows, we believe further rate declines, and hence the downside price risk for swaps, is relatively limited.

In contrast, in a rising rate scenario, swap rates have no such bound.

As a result, we believe the price profile of swaps at these low rate levels is somewhat similar to the price profile of an option.

With limited price declines and a falling rate scenario, and unlimited price gains in a rising rate scenario.

Given this asymmetric profile, we preferred pay fixed swaps over other hedged instruments last quarter.

As a result, we added $12 billion of new pay fixed swaps during the quarter, bringing the total swap portfolio to $48.6 billion at quarter end.

These new swaps had an average term of 6.7 years.

Given the addition of these swaps, the average maturity of our swap portfolio increased to 4.3 years at quarter end from 3.9 years the prior quarter.

On slide 13, we break out our other hedge position, namely, our treasury and mortgage hedges, in greater detail.

Taken together, our total hedge portfolio covered 91% of our debt balance at quarter end.

On average during the quarter, our hedge portfolio covered 88% of our debt balance, up significantly from 78% in the first quarter.

We chose to operate with a higher hedge ratio last quarter because we believed it was appropriate, given the confidence that we have in the performance of our assets in a falling rate environment.

And given the market conditions we faced.

As portfolio composition and market conditions change, we will adjust our hedge ratio accordingly.

Turning to slide 14, I'll review our duration gap information.

Due to the combination of falling interest rates and portfolio positioning, our duration gap shifted to negative 0.44 years at quarter end, from a positive duration gap of 0.41 years the prior quarter.

This shift in our duration gap was driven by changes in both our asset portfolio, as well as changes in our hedge portfolio.

The duration of our asset portfolio decreased to 2.7 years from 3.3 years, as interest rates fell during the quarter.

The duration of our hedge portfolio increased to 3.1 years from 2.9 years, as we added a significant amount of longer-term swaps to the portfolio.

Given the low interest rate environment, and the potential extension risk in our assets, we believe maintaining a negative duration gap and a higher hedge ratio was prudent from a risk management perspective.

At the current rate level, the duration of our asset portfolio is near its low point.

In contrast, if rates rise significantly, our asset duration could extend to about six years.

By maintaining a negative duration gap, we essentially pre-hedged some of that potential extension risk.

Additionally, in the current environment, we believe our assets will likely exhibit longer durations than those implied by our model.

With that, I will turn the call back over to Gary.

Thanks, Peter.

Now let's turn to slide 15 and take a quick look at the business economics.

Focusing on the two left-most columns, you can see that our net interest spread compressed to 162 basis points from 207 basis points at the end of Q1.

A little more than 50% of the decline came from the lower asset yield.

And the biggest factor here is the impact of our faster prepayment projections and what they had on premium amortization.

More specifically, the change in prepayment estimates accounted for at least 50% of the 25 basis point decline in asset yield.

Now, our cost of funds also increased during the quarter by 20 basis points.

This was a function of two main drivers.

First, repo rates were about 5 basis points higher.

Second, as Peter discussed, we made the conscious decision to increase our use of swaps and the duration of our hedges in response to record low interest rates.

The higher swap cost accounted for the remaining 15 basis points.

When you take our net interest spread, multiply by the leverage, and add back the asset yield, you get a gross ROE of around 15%, or a net ROE of 13.6%.

Remember, this ROE excludes realized gains and losses on our assets, and all of the impacts of our supplemental hedges, which also flow through the other income line item.

And yes, this ROE snapshot is lower than prior quarters.

But it is still compelling, especially against a backdrop of a materially lower risk position.

Now, if we turn to slide 16, I want to conclude by quickly summarizing how the Company was positioned at quarter end.

The bottom line is that we feel that our portfolio is tailor made for the current environment.

As Chris mentioned, almost 70% of our assets are backed by either lower loan balance or higher LTV loans originated under the HARP program.

Of the remainder, the majority are very low coupons that should remain relatively slow in the absence of a QE3.

If we get a QE3, the Feds' bid should drive valuations on those assets to very rich levels.

And we would likely sell those positions to the Fed if we had concerns around prepayment performance.

Lastly, as I said earlier, we are not betting on low rates or QE3.

But we can't ignore the realities of the market.

We have increased our hedges materially, which should help support the performance of the portfolio if interest rates rise for any reason.

Yes, these hedges cost money, but they are critical to our objective of producing attractive returns across a range of interest rate environments.

Moreover, if rates decline further due to weakness in the US economy, Europe, or Asia, it is highly unlikely that we would be over-hedged as mortgage performance should significantly outpace hedges.

So with that let me open up the call to questions.

(Operator Instructions)

Joel Houck, Wells Fargo.

The first question is, Gary, with the $1.2 billion capital raise in July, maybe talk about, I think it's implied that you guys continue to like the spec pool trade.

The economics on page 9, I think, suggests that.

The question is, did you add any more hedges or was the increase in swaps in the second quarter in anticipation of the equity raise?

To your last point first, Joel, we absolutely did add hedges with purchases that were made concurrent or around the same time as the equity raise.

The hedges that we put on in the second quarter were not in contemplation of doing an equity raise and pre-hedging purchases.

We were very comfortable with that position given the assets that we had at the time.

And with respect to new purchases, we continue to maintain essentially the same type of view with respect to hedging them that we had at quarter end.

In terms of the deployment of capital on the equity raises, is that similar types of themes that you guys have liked on a relative value basis?

Yes, in general.

Without going into the specifics.

We absolutely have added some specified collateral.

In the last week or two, the specified collateral has really taken another leg up in terms of price.

So our view is a little more subdued than it was at the end of the quarter given the example that we went over.

But we continue to view lower coupons, the combination of lower coupons and specified collateral as being the right place to be.

And when you're buying those assets, you also want to have a reasonable hedge ratio, and I think we're maintaining that same philosophy with respect to deploying capital.

Thank you very much.

That's great color.

Jason Weaver, Sterne Agee.

Gary, I was hoping you could give some more detail regarding the prepayment and duration assumptions on both the lower coupon and the HARP 30-year securities.

Maybe you can correct me here, but it seems that, aside from treasury starting to look more like the JJB curve, there are relatively few scenarios where you would expect much real refinancing demand for borrowers underlying these bonds.

I think you're right.

So, if you're looking at the, let's go with the HARP or lower loan balance securities that have demonstrated pretty good prepayment behavior over the last two or three years.

As Chris said, we don't expect them to pick up materially even in a further decline in interest rates.

Or, let's go to an extreme, even in a QE3 type of scenario.

So we think the cash flows are going to remain very solid on those assets, which is a key reason for being there.

The one thing I do want to stress, though, is you have to be cognizant.

And one of the real challenges, and something that we focus an incredible amount of attention on, is how do you hedge those positions?

You don't want to just sit there and say -- Okay, the prepayments are going to be slow so I should hedge them, in a sense, really long.

Because they're market prices and the way they trade will not really track that.

So I think you have to be more dynamic with respect to how you think about that hedging equation.

Now, with respect to the lower coupon fixed rate, I think we need to be a little careful here.

We do agree that if you're in the lowest coupons, the call risk or prepayment risk is relatively low.

But, when you start to get one coupon above that, let's say 30-year 3.5%s, or 30-year 4%s, where people have in the past said -- Those things can't prepay, I don't need to worry about prepayments on those types of assets, 15-year, 3%s.

We're already seeing those pick up, and there are many scenarios where those could be pretty fast.

So I think the reality is that mortgage rates are at record lows and a much higher percentage of borrowers now actually can execute on that refinancing option.

There are two categories.

The borrowers who have taken out a loan recently.

Let's call it the last two or three years.

They're generally good credit and generally have LTV.

The borrowers before that generally qualify under HARP 2.0, and clearly HARP 2.0 is working.

So, while I do agree with you, and we have tried to stress we feel good about our prepayment positioning, so to speak, and feel our assets will hold up, and that's a key reason why we're hedging them longer, I do want to be clear that I think there is more prepayment risk in the mortgage market generally than people are expecting.

Thank you, that's very helpful.

And one further question.

I think we've seen about six preferred deals in the space since years earlier this year.

It seems like they're getting pretty popular, for obvious reasons.

But putting the preferred market's liquidity aside for a moment, what would you say would be the optimal proportion of these within the capital structure, if you could raise as much as you needed to there?

I think it's very dependent on the market conditions and so forth.

It's dependent on how your common is trading, as well.

We felt very good about accessing the preferred market and making sure we knew where that execution was.

Generally speaking, however, preferred at this point hasn't been our choice in terms of trying to access it in a deep fashion.

I think it's very hard to say there's an optimal capital structure.

I think it's very dependent on market conditions and the execution on common.

Thanks very much, guys.

Bill Carcache, Nomura.

Gary, can you comment on your view of policy-related prepayment risk from here, particularly given the premiums being paid today?

You mentioned that it seems like HARP 2.0 is working.

Is that the extent of it?

Or is there a risk, do you think, that there could be additional policy risk on the prepayment front?

There's always policy risk.

And let's face it, if the economy takes another leg down, there will be more discussions, or whatever.

But big picture we feel that policy risk is probably as low as it's been over the last three years.

And the reason is because, to your point, HARP 2.0 is working.

There are things that can be done to make HARP 2.0 more effective.

You heard from Freddie Mac and FHFA two days ago that there are some changes in the works there.

There's the Menendez-Boxer bill that's being discussed that would do even more to reduce some of the frictions and make HARP 2.0 more effective.

There's also some discussion in that bill of moving the HARP eligibility date one year longer.

But let's take a step back.

There's still some policy risk out there.

There's still some discussion, even though FHFA has given its opinion on things like principal writedowns, which could have a minor impact, at most, on the prepayment picture.

But big picture, these things are very modest and incremental on the margin.

And they would apply to a very tiny percentage of our portfolio in aggregate.

So, big picture, when you think about what's even being discussed, and a lot of it's very unlikely that anything happens with it, it's very different from what was being discussed a year ago when people were talking about mailing a 4% mortgage rate to every homeowner who had paid their mortgage.

So I think people need to take a step back and say there's always risk on the policy front.

But if you look at what's being discussed at this point versus what was being discussed a year ago or a year before that, the changes are very incremental in nature.

And for most mortgage portfolios, they would affect a small percentage of holdings.

In the case of AGNC, even the Boxer bill would affect a very small percentage of our holdings.

That makes a lot of sense.

It's really helpful perspective.

Another question I had, you've crystallized for us how to think about what happens in different economic scenarios.

And if we focus on the QE3 scenario, you highlighted how it's not a Goldilocks scenario.

The curve would flatten and we'd see further spread compression.

But there would be a book value benefit.

If I could ask you to just look past a bit further out, past that initial book value benefit, with the premium in the industry's book relatively high, do you worry about the eventual loss of demand from the Fed?

And the extent to which there are no natural buyers out there for 30-year mortgages, and ultimately what that means for prices?

I'm just trying to get my arms around the risk that the big premiums being paid for agency today will somehow come back to haunt the industry at some point down the road.

If you could just help me in how to think about that.

First off, there's a real misconception out there that the only reason mortgage prices are where they are is because of government involvement in the mortgage market.

One thing that people need to focus on is that it is much safer and easier to hedge mortgages in this environment.

As Peter discussed with respect to, you could see it in the way we're hedging.

Mortgage investors have to be very concerned at times on paying on swaps or being short Treasuries or some of the things that we all do to hedge our interest rate risk.

Because if interest rates fall, mortgages prepay and don't keep up.

But as we've talked about, the risk of being overhedged, and the amount of prepayment risk in the new low coupons, is so much lower than it's been in the past.

And because of that, the risk-adjusted returns on mortgages are still very reasonable.

Even without a QE3.

And then if you go to the page in our presentation, page 9, where Chris went over the returns on the specified mortgages, yes, they're trading at high dollar prices.

But they're trading at high dollar prices because prepayments are reliable.

Once you take that out of the equation, then the returns are actually pretty reasonable.

So I think big picture, our view is that the mortgage market is supported by a number of different factors.

Yes, in a QE3 scenario there will be a time where certain coupons will be unattractive to most other people besides the Fed.

But we absolutely don't believe we've reached that point.

Great.

Thank you very much.

I'll jump back in the queue.

Thanks.

Daniel Furtado with Jefferies.

The first question is on this change in the premium, or the catch-up that happened in the quarter.

Can you help us understand?

Is this due to portfolio composition changes or was this movement due to your change in view of the future?

The catch-up component is not due to the composition of the portfolio.

It's due to raising our projected prepayment speeds.

If in the prior quarter we ran the portfolio at 9 CPR, and this quarter we're running the portfolio at 12.

And let's ignore the changing composition for a second because catch-up, Dan, really only applies to the existing portfolio, not the new part of the portfolio.

But it's related -- there are two ways that prepayment change affects your yields and your spread income.

The two ways are, first off, just because you're running it at a faster speed, the yield is lower.

And you're going forward, going to amortize premiums quicker.

That's the normal impact.

But you also have to, in a sense, catch up for the fact that you had been running the security slower in the past.

So there's a one-time or non-recurring change due to that increase in the prepayment speed, which is what we refer to as the catch-up component.

In the go-forward numbers is the running the security at the 12 CPR.

So that's more of a lasting impact.

But then there's the one-time adjustment where you actually account for the fact that you had been running it slower before.

Got you.

What I'm trying to reconcile is that your assumptions for the future for CPRs move up modestly.

But over the last couple months, we've seen CPRs move down pretty materially.

Are you just factoring this recent move in the 10-year or incorporating a potential QE3 in that forward outlook?

That's really what I'm trying to reconcile.

We've seen things improve but yet your expectation is for things to worsen from where they were a quarter ago.

One thing that we want to be very clear on with respect to the prepayment projections is that it is important when you do this to use a defined process.

So what we've said in the past, and it's absolutely still true with respect to the prepayment forecast, is that we rely on an outside provider, BlackRock Solutions.

And their model essentially reacts mostly to the change in interest rates.

So the flatter yield curve, and the considerably lower 10-year Treasury, and the higher mortgage prices are the key inputs to that model.

And then that model generates faster prepayment estimates.

Now, periodically, as new information comes in around the securities, where maybe they don't prepay as fast as the model had thought, then adjustments get made.

And then there's, we'll call it, a true up process.

But that's a slow-moving process realistically.

The key driver of that increase in prepayment speeds is purely the change in the interest rate environment.

And it is very much a model-driven type of result.

I want to be clear.

It's our responsibility to believe that the model is producing reasonable numbers.

And we do.

As we see the actual prepayments unfold, then we certainly factor that information in, in terms of projections going forward.

Great.

Thanks for that clarity.

More philosophically, I know that IOs are a touch dangerous should prepays pick up.

But at what point do IO securities start to make more sense for you?

I know you have a small amount on a notional basis now.

But is there an environment where you could see becoming much more aggressive on using IOs as a hedge?

Absolutely.

There are environments where we think IOs make a lot of sense.

I think one of the things that has led us to use less of them over the last year, let's say, because we had a relatively large position.

We had a decent-sized position at the beginning of 2011.

But most of the IO that's available to trade in the marketplace is in the HARP 2.0 eligible cohorts.

And we've had an opinion that that's not an area that you want to dedicate, where you want to take prepayment risk.

Now, we feel that as most people are probably aware, the higher coupons -- 5%s, 5.5%s, and 6%s, which are eligible under HARP 2.0, or most of them are eligible -- are now prepaying near or above 40 CPR.

So HARP 2.0 has clearly worked.

So we feel good about having avoided those cohorts.

With respect to IO off of lower coupons, if it were available in reasonable size, and if pricing made sense, it's certainly something that we would look at.

But in the HARP 2.0 universe, where most of the IO that trades is available, we've been bearish on it, so to speak.

Great, thanks for the time, Gary.

I appreciate it.

Bose George from KBW.

I was just wondering, your quarter-end spread, that 162 basis point number, I was just curious if you could give us an idea where that is right now.

We're not going to update any financial numbers inter-quarter.

But mortgages have certainly tightened, and payups have gone up.

But for the existing portfolio, the biggest moving parts were things that occurred during the second quarter.

In other words, if you look at the change in interest rates this quarter, and change in prepayment speeds, most, if any, impact there would have been felt in Q2.

Most of our change in our hedging framework occurred in Q2.

So if we talk about the drivers of spread compression, they're largely drivers that kicked in, in the second quarter, versus drivers that are kicking in, in the third quarter.

Okay, great.

That's helpful.

And then, actually, on your duration, I'm curious.

Your negative duration at the end of the second quarter, does that suggest that you'll be less benefited from QE3 than you would have been back in the first quarter?

It depends, is the short answer.

It's very possible that we will benefit less from a QE3.

So if a QE3 occurs and drives Treasury yield and swap yields lower, then, clearly, we will be hurt by having incremental hedges.

Again, we feel like our assets would perform very well in that scenario.

However, if QE3 occurs, and mortgages are the main beneficiary, but the equity market, let's say, improves, you could very easily see -- like what you saw with QE2 -- a situation where Treasury yields actually stay the same or go up.

In which case, then we wouldn't be impacted by the hedges.

Honestly, we're not calling for Treasury yields to back up in a QE3.

I think our mind set on the positioning with respect to hedges is one of -- let's be able to perform in either scenario.

The one thing we feel pretty strongly about is we don't want to have mortgages that are significantly exposed on the prepayment front, because there's really no good way to hedge that exposure.

We would rather have better-performing mortgages, and then use interest rate hedge.s We feel that's the right way to position the portfolio, and that's what you're seeing from us.

Okay, great.

Thanks a lot.

Arren Cyganovich with Evercore.

On the repo side, I'm just curious if you could talk a little bit about whether you are getting close to full capacity with your repo counterparties.

And how much additional upside do you have with your repo counterparties?

With respect to repo, we've certainly grown the number of counterparties, we've grown our repo balance.

I think what I would say is we feel we have sufficient repo.

We have some push-in or unused capacity.

But I do want to be clear.

Repo does not grow on trees.

And it's not like you can just push a button and have $20 billion of repo capacity just appear.

So we feel very good about repo.

We certainly feel that there are other counterparties out there that we can get repo from over time.

We feel that we can probably get additional capacity from some of our existing players.

But we'd also want to be very clear that it is not a limitless resource.

And that's certainly factored into all of our decision-making.

Okay.

On that same topic, what other options do you have if you intend to grow your balance sheet further, as you have the past couple of years?

The first thing is that we are cognizant of the fact that we're not going to maintain the same growth trajectory, given our current size.

The first piece is that people should just recognize that.

Second of all, I think there are a number of different options, which I don't really want to go through on this call, to over time expanding the funding options for agency mortgage securities.

There are a number of things that we're currently working on, on that front that we feel will provide benefits over time.

I think it's probably not appropriate for us to go into them on the call.

Okay.

Thank you.

Jasper Burch with Macquarie.

You spent a lot of time on the call talking about what risks you don't have and what risks you are hedged against.

I'm just wondering, on the current portfolio and on the new assets that you're putting on, what are the risks that you still see out there that are the most material or meaningful?

Or even the less likely risk that it's really hard to hedge against in this environment?

The two main risks obviously when we buy mortgages and we hedge those mortgages with interest rate swaps and other financial instruments, if mortgage spreads or prices decline without an offsetting move in our hedges, we will lose a fair amount of money in terms of the mark-to-market of the portfolio.

What people call mortgage basis risk is a key component.

Obviously there is prepayment risk.

I think we've talked at length about why we feel good about our position there.

There's liquidity risk and funding risk.

And I think we've talked about some of the things that we've done there.

And then lastly there's always negative convexity, which is the hardest thing to deal with when it comes to hedging a mortgage portfolio.

Which is, any big move in interest rates requires rebalancing and active management of the portfolio.

So that's something that we always keep track of.

There's obviously other risks.

You can read the Qs and Ks for additional risk factors, but those are some of the bigger ones.

Just on the negative convexity, putting on mortgages in this environment, lower yielding, higher basis price, it seems like that risk, if anything, is going up as you expand the portfolio.

Is this just based on a bet that we are going to continue in this lower interest rate environment for a long period, and then when rates do eventually rise it will be at a modest clip?

Let me start and then I'm going to hand it over to Peter.

I think that risk in terms of negative convexity and hedging is not going up.

Because you can afford to hedge the extension risk more so than you could in a normal environment.

In that, because rates are floored at zero or near zero, depending open how you want to think about it, your downside, when you put on edges right now, is very different from when if you put on a 10-year swap at 4% you could have lost 300 basis points.

Putting on a 10-year swap at 160 basis points, what's the Japan scenario?

Maybe it's 100 or 110 basis points, something like that for 10-year swaps.

That's a manageable kind of move.

So I think bigger picture, it's important to keep that dynamic in mind.

To your other point, to let Peter speak to you, we absolutely are not betting on the Japan scenario.

There would be plenty of other positions we'd have on.

Yes, we're certainly not betting on the fact that if interest rates rise, they're going to rise in a slow and measured way.

They could rise very rapidly.

That's really why I tried to point out that some of that extension risk in our portfolio has to be pre-hedged.

We don't want to rely 100% on our ability to rebalance the portfolio as interest rates are changing.

Buying in swaptions, for example, are a great way to pre-hedge some of that extension risk.

And we'll continue to focus on our swaption portfolio.

And using our duration gap, like I mentioned, positioning the portfolio at a negative about half a year, again, is a way to pre-hedge some of that extension risk.

So that when the rates rise, you're not forced to rebalance into a volatile market.

There could be a scenario where we would run even a larger negative duration gap than what we reported this last quarter.

Okay.

That's all very helpful.

If I could just ask one more.

In the opening remarks, you said that there's a tailwind to book value.

And I think we've all seen that, looking at the mortgage market.

How much of that do you think is driven by an expectation for QE3?

And how much more upside could there be in a QE3, if that occurs?

And then what's the downside if it doesn't?

Do you think the book value pulls back?

It's a tough question.

I will try to give you an answer, but I just want to be clear to people that there's no easy way to extract what percentage of QE3 is priced into the mortgage market.

Or what market participants think the odds are, and so forth.

My personal opinion is that mortgages should be tighter than people otherwise expect in this environment, for the reasons we've talked about.

The low rate environment, the ease in hedging them.

The fact that a 100-basis-point spread over a Treasury is a lot more valuable when the Treasury's at 100 basis points than it is at 300 basis points.

It's a much higher percentage of the aggregate yield.

So there's a number of reasons why mortgages should be pretty tight right now.

And I think that some investors are misreading that and they're assuming that this is all about QE3 is 90% priced in.

Our mind set is, and mortgages have done very well this quarter, but we still think that if you've got a large QE3 that involved large amounts of mortgage purchases from the Fed, that mortgages would end up being significantly tighter than where they are today.

Call it, 0.75 to 1 point tighter over time.

And, is there downside if QE3 gets priced out?

Definitely.

There's certainly players in the marketplace that view that as the trade, so to speak.

But we think it's considerably less than, we'll call it, the move that we would get if QE3 were to happen.

Okay, thank you, Gary and team, very much.

As always, you've given us a lot to think about.

Douglas Harter, Credit Suisse.

You've moved out your repo maturities pretty noticeably.

Do you have any plans to further extend those maturities?

I'm glad you asked.

We have made a very conscious effort to try to continue to extend our repo maturities, reduce our rollover risk.

We feel like it's the right thing to do, to continue to focus on that.

And we will do that.

We're finding good liquidity out, call it, in the one-, two-, and three-year sectors.

So we will continue to try to access that funding.

We think there's availability there and we think it's at a relatively attractive rate.

It won't be limitless, but it is something that we'll continue to focus on.

Could you just give us a sense as to how much more expensive a one-year repo is right now versus a 30-year, 90-day?

Yes.

Just round numbers, one- to three-month repo is probably in the 40 to 45 basis point range.

Six months, let's call it 50.

12 months maybe it's 60.

And then you can see from our table, out in the 25- to 36-month range, our cost is around 75 basis points.

So it is higher, certainly, than our overall portfolio average.

But at 25 basis points higher than the average, we think that's money really well spent.

Great, thank you.

Chris Donat with Sandler O'Neill.

Just one quick one for you, Gary.

You talked about the sustainability of the dividend over the near term.

I'm just seeing if I can pin you down on what you mean by near term, if that's quarters or years And then as a related function, the undistributed taxable income, should I think of that as a reserve account that will be there at some level in perpetuity?

Or is it an amount that you would be willing to draw down to zero?

Thanks.

Thanks for the question.

What I said was -- at least over the near term was the wording.

And I really can't be more specific than what I've said with respect to that.

Now, with respect to the undistributed taxable income, I think what's really important here is that our taxable income, let's just go back and talk about where it came from.

It came from the fact that our taxable income, over the last three years or so, has exceeded our dividend.

So we have basically accrued a lot of taxable income that has yet to be used to pay a dividend, but will be over time.

So I think the best way to think about that is something that materially reduces the risk, that taxable income will be a constraint on our dividend.

That doesn't mean that at some point we're just going to sit there and continue to pay this dividend or raise the dividend until that number goes to zero.

That's not the mind-set.

There are other factors that go into the dividend.

But I think the way you should think about undistributed taxable income is just the way I described it.

And as something that really limits the risk over the near term, that we're going to have a taxable income shortfall under a pretty wide range of scenarios.

The real factor, and what I want to stress, and what I said in the statement, was the reason we feel really good about the dividend is not only has taxable income exceeded our dividend, but clearly our true economic returns have exceeded our dividend, as well.

And that's why you've seen a consistent improvement in book value over time.

And it is the combination of those two factors that give us the confidence to make the statement that we made.

Got it.

Thank you.

There are no further questions.

The conference is now concluded.