AGNC Investment Corp (AGNCL) 2011 Q3 法說會逐字稿

Good morning.

My name is Terrence and I will be your conference operator today.

At this time I would like to welcome everyone to the AGNC Shareholder's Q3 2011 Conference Call.

All lines have been placed on mute to prevent any background noise.

After the speaker's remarks there will be a question and answer session.

(Operator Instructions).

Thank you.

I would now like to turn the call over to Katie Wisecarver of Investor Relations.

You may begin your conference.

Thanks Terrence.

Thank you for joining American Capital Agency's third quarter 2011 earnings call.

Before we begin, I would 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 fact, 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 forecasts 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 10-K dated February 25, 2011, and periodic reports filed with the Securities and Exchange Commission.

Copies are available on the SEC's website 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 website, and the telephone recording can be accessed through November 9 by dialing 855-859-2056.

The conference ID number is 16828279.

To view the Q3 slide presentation, turn to our website, AGNC.com, and click on the Q3 2011 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.

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

Thanks Katie, and thanks to all of you for joining us on the call and for your interest in AGNC.

What an interesting quarter.

Between the debt ceiling fiasco where our politicians did everything they could to create a crisis, to the US debt downgrade, the worsening of the European debt crisis, to surprises from the Fed, and growing prepayment risk.

Throw in the SEC concept release and Monday's HARP announcement and it is no wonder the REITs sector has been a touch volatile.

But with all of this we feel really good about AGNC's performance.

And perhaps more importantly, how it is positioned looking ahead.

Recall for a moment the prevailing view on interest rates at the end of the last quarter, the concern was that the growing US debt burden and the potential for a ratings downgrade would lead to a significant rise in interest rates.

Many very well respected investors placed large bets on that outcome.

Instead, we experienced the opposite; a significant decline in interest rates.

AGNC, on the other hand, bet on neither outcome and was positioned to protect shareholder value under either scenario.

The specific attributes of our portfolio that we highlight last quarter proved to be critical in positioning the company to continue to generate very attractive risk adjusted returns despite the flatter yield curve, record low rates, and even the recently announced changes to the HARP Program which we will discuss in detail in a few minutes.

With that as the introduction, I want to quickly review the highlights for the quarter on slides three and four as we have a lot to cover today.

As you will see we have also made a concerted effort to further enhance our disclosures related to our funding and counter party exposures.

And Peter will review them with you shortly.

First, GAAP net income totaled $1.39 per share.

Now if we exclude the $0.24 of other investment income, that technically leaves $1.15 per share.

However that number includes approximately $0.08 of non-recurring "catch-up" amortization which relates to a one-time adjustment to our historical amortization due to the increase in our prepayment estimates.

As such the $1.23 per share is probably the best surrogate for what analysts like to call core income.

The 13% CPR projection which drives our amortization and asset yields is approximately 50% faster than our actual Q3 CPR of 8%.

As such, if like other players in the industry we did not make forward projections to our prepayments and just used our 8% actual CPR our decline in net interest income would have been considerably smaller.

Alternatively, you may want to think of this as us having taken the margin compression we would expect to see over the next few quarters upfront versus waiting for actual speed increases to show up.

Now taxable net income increased significantly to $1.86 per average share as unrealized mark-to market losses on derivatives don't impact taxable earnings.

And taxable amortization was considerably lower than GAAP numbers.

To this point, taxable amortization was $34 million or $0.19 per share lower than GAAP amortization because prepayment projections do not impact taxable yields.

This is a key point to understand and you can see these numbers on slide 25.

Given these strong taxable earnings, our undistributed taxable income essentially doubled this quarter to $156 million or $0.85 per share.

Book value rose slightly to $26.90 per share from $26.76.

Our economic return, which is the combination of dividends plus the increase in book value, totaled 5.8% for the quarter or 22.9% on an annualized basis.

And as you can see on the next slide, our mortgage portfolio totaled $42 billion.

Leverage during the quarter was 7.9 times and was down slightly at quarter-end to 7.7.

Now turn to slide 5 and we can quickly look at what happened in the markets during Q3 The tables on the upper left show the significant rally and curve flattener we witnessed in both Treasury and swap rates over the past several months.

On the bottom left table you can see what happened to mortgage crisis.

Lower coupons rallied considerably more than higher coupons which makes sense given longer durations and great exposure to the backend of the curve.

This result was also impacted by the uncertainty surrounding the prepayment landscape which obviously included the concerns regarding the changes to GOC policy and the HARP program which obviously impacted higher coupons the most.

15-years on the top right show the similar pattern to what we saw in low to middle coupon 30-years, while ARMs generally underperformed on an organic prepayment fears.

Now let's turn to the next slide, slide 6, where we will summarize the new prepayment landscape.

A number of market participants have argued that prepayments will remain benign despite record low rates.

They have generally sided, the relatively low Refi index, tighter underwriting practices, capacity constraints, the high percentage of borrowers who may have trouble taking advantage of low rates.

And all of these arguments are actually valid.

But in our opinion they don't argue for benign prepayments.

They argue for very large prepayment differences between various types of mortgages.

As we stressed in the past, borrowers who took out mortgages between 2009 and 2011 are very good credits.

They are ready-qualified under tighter underwriting practices, generally have equity, and can refinance with few if any barriers.

This is what happened in 2010, and guess what, that is exactly what they are doing now.

Now, Monday's announcement from the FHFA did elevate prepayment risk on more seasoned mortgages as well.

And this adds a new element of risk to this segment of the market.

But before we get to this, let's turn to page 7 and look at the organic prepayment situation.

Now focus your attention on the far right portion of the two graphs.

Because these show the most recent prepayments from Fannie Mae which were released just about two weeks ago.

This is current data and it clearly demonstrates the point around differentiation between loan attributes.

On the graph on the left we show 2010 15-year 4% mortgages.

Look at the speeds on both the universe of 15-year 4%s and the TBAs which again is what you get delivered if you don't buy lower loan balance pools.

They already hit 30 CPR which exceeds where they were last year.

On the other hand, prepayments on the securities backed by lower loan balances between $85,000 and $110,000 have increased but are still around 12% CPR.

The exact trend can be seen on the right side of the page where we show the 30-year prepays.

The TBA or high loan balance 2010 30-year 4.5% spiked to 24 CPR while both the lower loan balance pools and the HARP loans remain very slow and well below 10% CPR.

The differences are pretty striking.

And what we take away from this data is that owning the right stuff is more critical than ever.

Now let's move to page 8 and we will discuss the incremental risk arising from the recently announced HARP changes.

First, I want to be clear that we strongly believe that these changes will have a negligible impact on AGNC.

Because as we discussed on our Q2 call, we were proactively selling our higher coupon seasoned paper and reducing IO exposure beginning in earnest during Q2.

As importantly, our core positions, newer low loan balance and existing HARP securities remained definitively outside of the scope of these changes.

So, we are even more confident after this announcement that our portfolio speeds will remain will behaved.

Now, with that said, we do believe the changes announced by the FHFA could materially impact prepayment speeds on seasoned fixed rate and adjustable rate mortgages.

While the elimination of the 125 LTV cap and the reduction in the fees charged by the GOCs should have a small impact, the significantly reduced reps and warrant exposure for originators and the reduction in the need for appraisals could be very important.

These changes could be material because appraisals slow down the processing of a refinance and significantly increase the exposure to the lender if the valuation of the house is later deemed to be wrong.

And additionally the apparent relaxation of the reps and warrants around a borrower's ability to pay also reduces originator's underwriting exposure.

So, given these issues and also the significant airtime that the program changes are receiving in the press which should raise borrower awareness, speeds of some of the most exposed seasoned cohorts could be above 40 CPR for some period of time.

But we won't know realistically the full story here until the first quarter of 2012 when actual speeds come in.

As such, we hope our investors agree that not having this exposure hanging over the portfolio is a meaningful positive.

Before I turn the call over to Chris I want to quickly highlight on slide 9 why prepayments are just so critical right now.

Conclusion from the tables at the bottom of the page is really obvious.

Faster prepayment speeds materially affect yields and ROEs.

The key point here is the magnitude of these differences based on quarter-end prices.

Now keep in mind that pools with favorable characteristics do trade at meaningful premiums to generic mortgages, but this table is based on TBA prices for simplicity and coverability.

Look at 30-year 4.5%s at speeds ranging from 10% to 40% CPR.

The yield is 3.48% at a 10 CPR, only 2.06% at a 30, or 1.19% at a 40 CPR.

Now as we saw in slide 7, this entire range of speeds is possible depending on what types of pools you own.

Now look at the impact for an investor like AGNC with assumed leverage of 8 times.

Gross ROE on a 30-year 4.5% can be as high as 24.5% at a 10% CPR and as low as 4% at a 40% CPR.

This is really why we obsess about asset selection and this slide is relevant for both our existing portfolio and for new purchases.

And the bottom line is that new purchases can still produce very attractive returns if you can control for prepayments.

And this is even after accounting for the payoffs that you have to pay.

At this point I am going to ask Chris to review how AGNC's portfolio is positioned against this backdrop.

Thank you, Gary.

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

At a high level the portfolio is similar to where we ended the second quarter with 15-year representing approximately 53%, and 30-year representing 38%.

Recall in our second quarter discussion that the vast majority of these positions were backed by loans with favorable prepayment characteristics.

And so we were extremely well positioned for the rally that occurred during the third quarter.

That being said, we did make some meaningful adjustments during the quarter in light of renewed focus at all levels within the government on making improvements to the HARP Program.

What is particularly relevant on this slide is that our average prepayment speeds continue to be very well contained despite the rally in rates and subsequent material increase in prepayments on more generic MBS.

For example, the entire cohort of 2010 originated Fannie Mae's 15-year 3.5%s paid over 27 CPR for the October factor release.

Again, that is 15-year 3.5%s at 27 CPR.

That is 125% increase versus the prior month.

The actual prepayment speed on our portfolio came in at just 9% CPR for the most recent October factor release and averaged just 8% CPR during the quarter.

Let's turn to the next slide for more detail on why we have experienced such favorable performance relative to more generic MBS.

As Gary mentioned earlier, prepayments are going to drive performance and we were worried about two forms of prepayment risk, organic prepayment risk and policy risk.

Given today's record low interest rate environment, organic prepayment risk, or in other words traditional refinancing risk driven by record low mortgage rates is very high on loans originated over the past two years.

At the same time we had all levels of government evaluating enhancements to improve the effectiveness of the HARP program and hence introducing significant policy risk on the HARP eligible cohorts.

On slide 11 we have given you more detail on the breakdown of the composition of our holdings.

Within the 15-year which represents 53% of our portfolio, the vast majority, more than 90% are backed by loans with favorable prepayment characteristics.

Over 88% of these pools are backed by loans with lower loan balances.

The lower loan balance category is comprised of pools backed by loans with original loan balances less than or equal to $150,000 and have a weighted average original loan size of $104,000.

I also want to point out that within the $1.6 billion "other" category, approximately $640 million are pools backed by loans with original loan balances less than or equal to $175,000.

Now, within the 30-year sector, more than 86% of our holdings are backed by pools with loans that were either originated through the HARP program or have loan balances less than or equal to $150,000.

This component of the position increased during the quarter as we continued to sell seasoned higher coupons and other generic MBS as we had done during the second quarter.

With respect to policy risk, it was our opinion during the quarter that the most exposed category to the HARP program enhancements were higher coupon pools issued prior to 2009, of which less than 5% of our portfolio was comprised of single family fixed rate or ARM pools issued prior to 2009.

While seasoned higher coupon pools have provided reasonable performance over the past couple of years given an inefficient HARP program, our opinion was that the risk and return profile was no longer compelling in light of the renewed government focus on improving the program's effectiveness and relatively full evaluations.

The next slide summarizes the positioning of the portfolio versus segments of the overall mortgage universe.

It also overlays a heat map that outlines our views on prepayment risk levels for these categories.

While I am not going to review the slide in detail by comparing the last two columns on the right side of the page.

Investors can get a very good feel for how different AGNC's portfolio is relative to the mortgage universe.

The table demonstrates what should be pretty obvious at this point.

AGNC is extremely overweighed lower loan balance and HARP securities and AGNC has very little exposure to new generic securities or the seasoned universe.

With that, I will turn the call over to our Chief Risk Officer, Peter Federico.

Thanks Chris.

Protecting our capital over a wide range of interest rate scenarios is our primary objective.

Careful asset selection and prudent hedging is the foundation of that framework.

As Gary mentioned in his opening remarks, this means that we do not make significant bets on particular interest rate outcomes.

Today I would like to discuss three important topics that have been the subject of increased focus this past quarter.

These topics include liquidity management, counter party credit risk management, and lastly changes to our swap portfolio.

With that introduction, please turn to the financing summary provided on slide 13.

During the quarter we undertook several important actions that significantly enhanced our liquidity position.

First, our repo balance increased to $38.8 billion at quarter-end from $33.5 billion the prior quarter.

More importantly, we significantly reduced rollover risk by lengthening the contractual term of our repo funding.

In percentage terms, the share of our repo funding with maturities greater than 3 months increased to approximately 20% from about 12% the last quarter.

Given the longer term of our repo funding, and slightly higher repo rates in general, our average funding cost increased slightly to 28 basis points from 23 basis points last quarter.

Our ability to extend the maturity of our repo funding was a direct result of another important decision we took during the quarter which related to hedge accounting.

At the beginning of the quarter we decided to not designate new pay fix swaps as cash flow hedges for accounting purposes.

Additionally at the end of the quarter we decided to discontinue hedge accounting for the remainder of our pay fix swap positions.

Without the funding restrictions associated with hedge accounting which required us to precisely align the term of our repo funding with the swap reset dates, we were able to begin to move a greater share of our funding into 3, 6, 9 and 12 months maturities.

Going forward we will continue to enhance our liquidity profile by opportunistically accessing funding across the full scope of the repo funding spectrum.

As of quarter-end the contractual original dates to maturity of our repo book was 57 days.

As of October 21, the original dates to maturity had extended to about 70 days.

Our decision to discontinue hedge accounting will also have significant impact on our financial statements, most notably going forward fair value changes associated with our swap contracts will be recorded in the income statement through other income.

As a result we expect our GAAP earnings to be more volatile and more sensitive to interest rate changes.

It is very important to understand that this increase in earnings volatility is not reflective of any change to our true economic risk profile and that book value remains the best measure of our aggregate mark-to-market performance.

I would now like to turn to slide 14 and discuss counter party credit risk management.

Slide 14 provides new information regarding our counter party exposure that we felt was important to share given the heightened sensitivity to this area in recent months.

First, the table on the bottom left of the page shows the dispersion of our repo counter parties by geographic region and by percent of funding.

During the quarter we took actions to ensure prudent counter party and geographic diversification.

We also increased the number of our repo counter parties to 29, from 26 last quarter.

We continue to diligently manage and monitor excess funding capacity both in aggregate and by counter party.

The table on the right shows the economic exposure we have to each counter party.

The middle column of the table shows our exposure expressed as a percent of our equity.

We compute these exposures daily taking into account the actual repo haircut as well as daily market value fluctuations.

Our largest single counter party as of quarter-end was less than 4% of our equity.

And our top five exposures together totaled less than 15% of our equity and we are well diversified geographically.

A final thought on funding and counter party risk management.

Despite increased concerns regarding the European debt crisis, at no time during the quarter did we experience a disruption to our repo funding.

At times during the quarter repo levels did increase, but these episodes were short-lived and the increases were measured in single digit basis points.

At no time during the quarter did we experience a reduction in liquidity or an increase in margin requirements.

I would now like to turn to slide 15 and discuss our derivatives portfolio.

As shown in the table our pay fixed swap balance increased 22% to approximately $27 billion at the end of the quarter, up from $22 billion last quarter.

Given the increase in our swap book, about 70% of our repo funding was economically hedged with pay fixed swaps at quarter end.

In addition, given the historically low level of swap rates observed during the quarter, we concentrated our swap activity on intermediate and long-term swaps.

Despite our focus on longer maturities, the average pay rate on our swap book decreased to 1.58% in the quarter from 1.69% last quarter.

Our swap shift balance decreased to $3.3 billion during the quarter from about $4 billion last quarter.

Given the rate of decline during the quarter, the market value of the portfolio dropped to around $5 million from about $36 million last quarter.

Looking ahead with interest rates at such low absolute levels, we view the risk of maintaining a relatively longer duration hedge portfolio as being significantly diminished.

Today, the risk of over-hedging mortgage assets with intermediate term swaps is significantly lower given the fact that swap rates are bounded by a zero floor.

Now let's turn to slide 16 and 17 which cover our duration gap and interest rate sensitivity.

On slide 16 you can see we had a very small duration gap of negative 0.3 years or about a half of a month at quarter-end.

Our duration gaps declined during the quarter from a positive duration gap of about 7 months last quarter.

This decline was due to the rally in rates and the inter-quarter activity.

Duration gap can be a useful measure of comparing relative risk levels from one time period to another.

But as highlighted on the slide, duration gap does not take into account the negative convexity of mortgage assets.

For that reason on slide 17 we provided you the same interest rate sensitivity that we typically provide in our 10-Q.

The table on slide 17 shows the estimated change in value of our portfolio for 50 and 100 basis point rate shocks.

The change in value is expressed both as a percent of market value of our assets in the middle column and as a percent of our equity in the right hand column.

Our interest rate sensitivity at quarter-end provides a good segue into the last area I will cover which is our goal of protecting book value, a topic I briefly mentioned at the beginning of my remarks.

Slide 18 shows the quarterly change in our book value over the last 11 quarters plotted against a graph of the 10-year Treasury yields.

As you can see, in 10 out of the last 11 quarters our book value increased.

This consistent performance is particularly meaningful given the extreme interest rate and prepayment episodes experienced during this time period.

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

Thanks Peter.

So let's turn to page 19 and quickly review the quarter-end numbers.

First, our asset yield declined from 3.45% to 3.18%.

This was driven, again, predominantly by the increase in our projected CPRs and the changes to the composition of our portfolio.

Our cost of funds also increased 15 basis points to 1.24% as of September 30.

And repo rates increased and we increased the percentage of swaps to close to 70% of our repo balance.

This dropped our net interest spread as of quarter end to 194 basis points.

And after adjusting for leveraging and adding back the asset yield, this provides a gross ROE of just over 18% or net ROE around 16.5%.

And again, this is without regard to any potential benefit from other income.

Now let's turn to page 20 so I can conclude by giving you a little more transparency into what we think the future could hold.

First, we remain really optimistic that we can continue to produce strong returns for our shareholders despite lower rates, flatter curve, and a more challenging prepayment environment.

In fact, in the absence of negative surprises to prepayments, significant changes to interest rates, or material changes to the portfolio, we expect asset yields to be relatively stable.

And remember, we are already projecting faster prepayment speeds and we would need prepayment picture to worsen beyond these assumptions to negatively impact our REOs.

With the HARP changes already announced, this risk is considerably lower given the composition of our holdings.

Alternatively, we would have upside in the event that actual speeds are below our forecast.

Now with respect to our cost of funds, we have significantly increased our percentage of hedges and materially reduced our duration gap.

We have also accounted for higher repo rates than we currently see.

So, lastly, new purchases are still attractive or just have cheapened versus swaps as risk premiums related to prepayments have increased.

In conclusion, we remain optimistic that AGNC's portfolio will continue to perform very well in this new interest rate environment.

So with that, let me ask the operator to open up the call for questions.

(Operator Instructions).

Our first question comes from the line of Bose George with KBW.

Good morning.

I had a couple of questions.

First I just wanted to see or check what kind of premiums do you need to pay for the prepayment protected funds that you guys have relative to TBAs?

It's very dependent on the coupon.

And whether it is 15-year or 30-year.

So, in the last we will say month or two, those premiums, and in particular very recently after the prepayment release showed the release in the newer coupons, the premiums have gone up.

But, let's say for something typical they are close to a point, maybe a little more on some, less on some.

So that point in yield terms is something like 25 basis points.

But if you go back to that ROE slide on page 9, 25 basis points is still a very small kind of increase or difference in CPR.

So, what you can see if you layer that into those charts is that the breakeven prepayment for something with a 1 point payout is maybe 4 CPR, depending on the coupon.

Great.

And then switching to HARP funds, and given the change could create a new flow of HARP funds, is that something where we could see that continue to be a big part of your portfolio?

Great question.

We are extremely happy with the performance of the HARP funds.

And given the recent release, the HARP securities are still not refinanceable and probably won't be now for the future, or won't be given what FHFA has said.

As such, I mean when you look at the prepayment fees, when you look at the fact that there are no options for most of these borrowers, and they already have pretty low rates, this is a great security to own.

Now one thing you have to keep in mind there is that the new coupons that are going to produced are lower.

That doesn't mean they are not attractive but the advantage of having them in 4.5% and 5% coupons is much greater than the advantage would be in a 3.5% or 4% coupon.

So the short answer is it is our favorite position.

We feel better about it now.

We are going to continue to take advantage of it, but the change in coupons is material to the equation.

Great.

Thanks a lot.

Thank you, Bose.

The next question comes from the line of Mike Taiano with Sandler O'Neill.

Thanks for taking the question.

So, it seems as if this quarter you took a number of actions to, I don't know if you want to use the term de-risk, but it looks like in terms of your hedging strategy and extending out repos that you lowered the risk profile during the quarter.

It is fair to say at this point how you have kind of structured the portfolio that you feel like you can take on more leverage, particularly with the hedges that you have put on in the quarter?

The short answer would be yes.

The portfolio is sort of lower risk, both in terms of prepayments, hedges and so forth.

In a sense running an equal risk position today could afford materially higher leverage.

I think the thing you have to keep in mind is some of the uncertainty that we are seeing overseas and so forth.

That is a consideration but if you got past that, not that we are past the uncertainty around the changes to the HARP Program, higher leverage is a realistic future outcome, and not from the perspective of pushing it around returns, just from a perspective of the risk equation if you are appropriately hedged and if you have prepayment protected assets is something you are supposed to think about.

And then just a follow up to the previous question.

It is fair to say, and you guys always talk about relative value, the relative value equation here on the HARPs and the lower balance assets, do you still feel similar to as you did going into the quarter on a relative value basis on those relative to say hybrid ARMs given the price differential at this point?

Honestly in terms of the relative value, even though the premiums for both low loan balance and HARP loans have gone up, and they have gone up again in the last couple of days as they have essentially gotten scoped out of the HARP Program.

There is still value for a long-term investor in those products although even with the higher prices.

Honestly in a fast prepayment environment like today, hybrid ARMs are not high on our list of considerations.

We do feel comfortable and would feel comfortable buying more generic lower coupon, 15-year or 30-year, where coupons may be low enough this time around, where some of that protection isn't necessary.

But I would just say at this point, hybrid ARMs, this is a prepayment environment, we have talked about that.

Hybrid ARMs really don't give you a place to hide on that front.

So they are really not on the kind of radar screen at this point.

Great, that's helpful.

Thanks guys.

Our next question comes from the line of Jason Weaver with Sterne Agee.

Good morning.

Congrats on another solid quarter.

Just a couple of questions.

What would you estimate the drag was on net interest income from those repositioning out of the high coupon paper and what are you expecting from your run rate spread going forward?

So, two things.

I don't have the number for you on the first one, what the drag was.

But if you look at the aggregate moves, right, in terms of three things effected our net interest spread kind of materially.

One would be higher cost of funds.

Another would be the faster, again, projected prepayment speeds because our actual prepayments did slow down.

And the last would be repositioning of the portfolio.

I don't have the breakdown of those but the reality is that between the two relating to asset yields, we think a prepayment one was the bigger impact.

The prepayment one being faster projected speeds.

The one thing I will say around your second question which is where you asked me to project out our margins going forward is I will just reiterate what I said on the last slide.

I mean you know the components of margin, their funding costs, and asset yields.

We are already projecting our asset yields for a material increase in prepayments, despite the fact that we have very good collateral and that is just out of model given the significant drop in interest rates.

So, given the fact that that is projected, we are not waiting for actual speeds to show up.

And remember that higher CPR is a lifetime CPR.

It doesn't mean you can't have a few prints and it actually incorporates some initial prints well above the 13 [deal number].

So keep all of that stuff in mind as you think about that going forward, but again, we don't tend to give guidance.

Fair enough.

And just a little bit further clarification on that.

You touched on it.

The 2.50 assumed duration on the fixed rate assets that you show on slide 16, does that take into account the higher model durations of lower loan balance and HARP funds?

It does.

On the other hand we are very careful not to, and this is a really important question.

We don't use what we would call theoretical model durations for HARP funds and low loan balance.

And the reason is they are too long and they don't...they are not informed by the way the securities are priced and the way they trade.

In a sense the model assumes that they trade at their full theoretical valuations.

They don't.

The model also assumes that they can trade below the price of a TBA or generic mortgage if interest rates go up.

Well they can't because they are deliverable and you would just sell it into the TBA.

So for that reason, while we use longer durations versus TBAs for the low loan balance and the HARP loans, we are very, very careful to manage those to kind of realistic market pricing assumptions and we absolutely do not blindly follow our model.

And then just finally, it seems like the positioning has been particularly well suited for the speed bumps that we have encountered in the market this year.

Now that we are at this point, where does your attention turn to on the risk side now that you are less exposed to HARP 2.0?

I think it is way too early to say that prepayments are kind of behind us.

Look, the bottom line is where any low interest rate environment where prepayments are going to dominate ROEs unless something changes.

Now, two things can change, right?

We can have a further decline in interest rates in which case prepayments become more important.

Or, I mean obviously we can stay here or we can go up in rates.

And if we go up in rates prepayments will obviously be less of an issue.

But because we have the confidence in our portfolio and how it can perform if interest rates fall, it allows us to be a little more aggressive so to speak on hedging at this point which is what Peter went over.

Which is why you are seeing hedge ratios.

So when you put those two pieces together, we feel the threat so to speak of faster prepayments we have contained, you know another risk is higher interest rates, so we have taken steps to contain that.

And our mindset, again, is we don't know which way interest rates are going to go.

We don't know how Europe is going to resolve itself and everything the Fed is going to do and so forth.

We are not kidding ourselves.

We are trying to be able to make money and protect our book value in either of those outcomes.

All right, thank you very much.

Our next question comes from the line of Mike Widner with Stifel Nicolaus.

Good morning guys.

I was just wondering if you could maybe elaborate a little more on your thoughts as far as leverage and the portfolio size go.

And specifically, I guess, you finished the quarter at 7.7 times.

You averaged 7.9 during the quarter and both of those are a little lower than where you were on average last year.

And you also talked about with the increase in the swaps book, you potentially got the chance to add assets and you sort of already have the hedges in position in some regard.

So, in putting all of that together, I don't know, just help me reconcile the lower leverage versus the quarter average against the other things you see out there.

Honestly I think that the lower leverage versus quarter average, the difference is not significant.

It is not that significant.

But big picture, look, there was a lot of uncertainty as I kind of started the discussion with this quarter.

Just everything that was going on and, look, we are not, we feel like we do a reasonable job on hedging the portfolio for different outcomes.

But, we recognize the volatility.

And I think that was a key factor in decisioning with respect to leverage.

Let me follow up on leverage which was, I think, the first part of your question.

And I want to stress something that people don't focus on enough which is there is a other massive benefit for having a slow prepaying portfolio.

And that is that it significantly improves your liquidity.

I am going to use an extreme example, if you run risk of having prepayments in the 40% or 50% area, on a monthly basis that means 4% or 5% of your portfolio is factoring down.

And given the delay between when those factors are released when you get the money from the GOCs.

That 4% or 5% is essentially an incremental haircut.

So, if you have to allot for 4% prepayments and you had a 4% haircut to begin with, you essentially get double the haircut with half coming from the posting a margin or the initial haircut, and half coming from the prepayment issue.

And so another key advantage for AGNC is significantly improved, like liquidity positions, as a function of not having to allocate a lot of room in a sense for prepayments.

When you think about the HARP Program, when you think about low loan balance, we can expect kind of muted changes going forward and that helps us a lot in that equation.

Just following up on that, again, another sort of reason between having the heavy swaps position and having the slower prepays which basically affords you great liquidity, both suggest that there is greater comfort or cushioning at least for sort of higher portfolio size, higher leverage.

And that is really just what I am trying to get a feel for, directionally speaking, you headed down at the end of the quarter versus where you had been and where you were during the quarter.

And is that kind of a trend line we should extrapolate?

Is this a current level we should extrapolate from?

Or should we assume you are going back to something close to 8 times, or maybe north of 8 times which is certainly territory you have been in quite a bit before?

What I would say is it is going to depend on market conditions.

But I want to just reiterate what I said is that in the absence of kind of real volatility and uncertainty, the more kind of benign hedging environment, the things Peter mentioned, kind of the zero floor and the lower risk of being over-hedged, coupled with confidence on the prepayment side given our portfolio, all of those issues lend themselves to us being comfortable with noticeably greater leverage than where we are today.

On the other side you have got the EU meeting this afternoon and a fair amount of volatility.

Again, we are not worried about counter party risk.

We are not worried about availability, up repo or any of those.

Those are not the concerns.

The concerns are just big picture, the volatility.

So, if we get more visibility around kind of volatility then higher leverage is definitely something we will consider.

The other factor is how cheap mortgages are and what we think about the risk adjusted returns.

We are not going to layer on more mortgages if we feel that they don't offer significant value and if they are not going to produce kind of very good risk-adjusted returns.

Right now that isn't a constraint.

Great.

Okay, so let me ask you one other question on the swaps.

At 70% swaps more or less, and that is quite a bit higher than where you have traditionally been, higher than where most of the group has been.

And so the more normal level is say 55%, just kind of picking that arbitrarily, where you guys are today at 70% by my math presents roughly a $0.10 to $0.12 drag on earnings, all else being equal, just from having that sort of excess level of swaps.

And I kind of juxtapose that against your comments early on that, some peers and some other smart investors out there made big bets on interest rates going up a quarter ago and that didn't work out real well for them.

And so given the additional level of swaps is inherently a drag on earnings, are you effectively making a bet at this point that rates go up.

And, again, you also I think talked appropriately, proudly about the ability of you guys to raise book value consistently over the past couple of years.

And as we look around at the world, and one way to sort of increase your likelihood of that happening is effectively making a bet on interest rates one way or another.

So, how should I think about all of those issues with regard to the hedges.

Turn to page 17 which Peter reviewed and what that shows is, an again models are anything but perfect.

We have talked about it.

But this just kind of gives you, and again we use BlackRock Solutions, that is our base risk management system, so these come out of that system.

Basically what this shows you is in a sense we are as "flat" as we could have been at the end of the quarter.

So you shouldn't read into that in any way, shape or form that we were, that we are making a bet, so to speak.

We just feel that given how low swap rates are and were, it was a prudent move and there are, we take risks when we feel that we are getting paid a lot for them.

And to be perfectly honest, 10 basis points in spread right now is irrelevant compared to the prepayment equation, right?

Look at slide 9 again.

Having the right securities is where 50, 75, 100 basis points.

The 10 basis points in feeling better about the hedge environment is just not high on our priorities.

And this Peter.

I would just add one thing to that, a number I didn't share in my prepared remarks, but the marginal cost of swaps that we entered into was just a little bit north of 1%.

So that compares pretty favorably to the 1.6% of the existing swap books.

The marginal cost of hedging right now is really low.

I get that.

Quick comment on page 17.

Does anybody seriously entertain the notion that we are going to have a parallel upward shift in the yield curve?

And it is kind of impossible to have a downward parallel shift in the yield curve, so I am not sure how useful.

I understand it is the boiler plate template that everyone uses but I am not sure that is a very good indication of real interest rate sensitivity, but I certainly appreciate the comments there.

I appreciate your comment.

And that is one of the reasons why we talked about the fact that for our swap book we essentially have kind of a zero bound on rates.

In fact swap rates could go to zero, but more practically and I think to your point, if we have a significant rally, they will rally 50 basis points or we could have 100 basis points, so there is certainly asymmetry in our interest rate risk profile just generically right now.

And I appreciate your comment about the fact that, yes, it is not like we are going to get a parallel shift but the downward scenario would be significantly less than the upright scenario right now.

Great.

And one final one if I may.

I'm looking at the income statement.

I'm wondering if you could, particularly the $263 million on sales, that was a surprisingly large number to me just given as I look at the templates that you provide on what is in the portfolio.

It looks like there was a lot more MBS sales in the quarter than I can sort of derive from looking at what changed.

I'm just wondering if you could maybe comment on the dollar value of, whether fair value or principal amount or whatever of MBS that you sold in the quarter.

I'm trying to reconcile how to get to a number that large.

Sales volume was a little over $10 million, but there are a couple of drivers of that.

So obviously a chunk of it was what we would call the seasoned higher coupons, in particular 5% coupons.

You will notice that we had a lot of them on our prior chart in the appendix and there were a lot less of them in the age of the 5%s is considerably newer.

That was a component of it.

Another component just as an FYI was we did view there to be some risk in terms of the HARP Program changing and borrowers being allowed to re-HARP.

We didn't think it was a smart decision.

But we had to consider that.

So one of the ways as an example we managed our exposure to that program risk which now luckily is now gone is that we sold our highest payouts, we sold as an example some of the HARP loans that were trading with payouts close to 2 points at different points during the quarter and replaced those with other HARP loans that had lower coupons and lower payouts that were well less than a point.

And so in doing that we could own more of them with actually less payout risk if the program had changed.

So there are other kind of transactions like that that you won't see in the numbers that drive, that have a big impact on those numbers.

I think at this point we probably need to flip over to the next question because I know we have a number of people in the queue.

I thought I was the most important, but thanks, I appreciate all the comments.

Thank you.

Our next question comes from the line of Steve DeLaney with JMP Securities.

Thank you.

Gary, great job in a tough quarter man.

And especially on growing the book value there.

Good work.

Thanks.

I will try to be quick on these because I know you are dragging on here for almost an hour..

So first on the undistributed taxable.

Obviously we are in a world of flatter curve and faster speeds, so spreads are tighter.

You have disclosed that and given us your 1.94 sort of spot spread.

So it forces to kind of look at the cushion, I call it your cookie jar when we think about the dividend going forward.

So, $156 million or $0.85.

I just want to ask you how you think we should think about these GAAP tax differences, and I guess it is $80 million or $0.40 or so of unrealized derivative losses and the differences in premium amortization.

When we look at the cookie jar should we look at it as $0.85 less $0.40 of taxable expense coming in the next couple of quarters as those items come through the system.

It seems you have a cushion but I'm not sure we should value the cushion at $0.85.

Any comment on the way I am looking at that?

Well, what I can comment on is kind of the GAAP tax differences which is probably the best way to think about undistributed taxable income.

So, a key point, and this is the amortization difference.

There is a pretty big amortization difference between GAAP, which again, this gets to the point we were trying to stress earlier on which was that GAAP, we are using a project number which is well above where things are now and where purchase assumptions were.

And for that reason GAAP amortization including this "catch-up" component which is a one-time thing was above, materially above the tax number.

So, first off, if you thought about a taxable run rate, it is going to be higher than the GAAP run rate and that $0.19 difference between the two should give you some insight into the difference of the two different amortization schedules.

So that is the first key point.

Then, with respect to the timing differences in unrealized gains and losses that make up the other component, the $0.45 of GAAP versus tax, there are two sides of that.

The first of there were timing differences in the previous quarter, going into the quarter, which actually I believe that number was around $0.15 where in a sense the first $0.15 of your $0.45 actually was from Q2 differences.

And so there is really $0.30 for the most part "out there" so to speak that goes in the other direction.

Is that right, Bernie?

So $25 million was the --

I'm sorry.

Yes, I'm doing that in cents, not in dollars, so that is a good point.

But, so $30 million going out there and actually where you can almost see the bulk of that, Steve, is in the Swaption numbers.

So essentially what happens is we mark-to-market the Swaptions which as Peter told you the mark-to-market of the Swaptions is not exactly a large number at this point in time.

On the other hand for taxable purposes you are going to recognize the difference between what you paid for the Swaption and if you exercise it, then you will get that value.

If you don't exercise it and it expires, then you will take that loss.

And we've got Swaptions going out a couple of years.

So that really gives you kind of a good feel for the bulk of the recognition of those differences.

But first off, it is just not nearly as big of a number as you might think given the component of the amortization number and also given the fact that it is both forward looking and backward looking in terms of timing differences.

That's helpful, Gary, especially the fact that the premium amortization difference is actually going to be an ongoing thing until such time that the actual might catch up with the projected.

We will spin all of that through in terms of trying to make your modeling assumptions.

And then one final thing, obviously the company has grown, a lot of hires, and obviously the management piece is going to be just a function of equity.

But we do know that the G&A expenses basically doubled from first quarter to third quarter, from $2.6 to $5.8.

What is really driving that?

Is comp reimbursements stuff?

What is going on there?

It is actually more in terms of things like prime brokerage, it is varying settlement patterns.

There is no compensation component.

That is kind of disallowed.

There is no way there could be a compensation component to that number.

So it is literally paid fees around settlements and so forth.

And BlackRock Solutions, different services that we use, data that we pay for.

It is things like that.

Maybe consulting services and those kinds of issues.

So transactional and obviously in a quarter where you have a lot of activity in and out that number might trip higher than in a quarter where the portfolio saw less turnover.

Absolutely.

That is a very good way to think of it.

It is probably most affected by, in a sense, transactional volume as opposed to other aspects.

Very helpful.

All right, thanks for the time.

Thank you, Steve.

Appreciate it.

The next question comes from the line of Joel Houck with Wells Fargo.

Okay, thanks.

In looking at page 18, obviously there is clearly a relationship between benign prepayments being environments versus high refinance risk in terms of your impact on book value.

But given that book value tends to do better in a higher rate environment, how does that square with the disclosure on page 17.

I understand that 17 is just a model, but if you are really going to lose 7% of NAV for 100 basis point increase in rates, that would seem at odds with past experience.

That is not the case.

Is there perhaps more realistic disclosure you can put out in terms of forward interest rate sensitivity?

Look, what we have to do here is use kind of a consistent model method.

Look, the reality is mortgages are negatively convex and big moves in interest rates, we lose money on either side.

You are absolutely right in saying that actual results in a sense are going to vary definitively from the model.

And so that is core.

What is interesting, and I think the point that you are in a sense trying to tease out, is generally when interest rates go up, especially from a period where prepayment risk premiums are large, then mortgages in a sense tend to perform better than a model would think because those risk premiums sort of come out of the market as overall prepayments come down.

And that is a lot of what we saw in Q4 and Q1 of 2010, early 2011, and what you certainly could see if rates were to go up quickly over the next year or so.

Conversely, when interest rates fall you have the opposite, risk premiums around prepayments go up and so mortgages tend to do worse than a model would actually project.

And this is, again, in and above just the absolute projections of prepayments, even assuming the model was perfect on those projections.

It is really that the market's tolerance for risk and big rallies goes down and vice versa as they sell off it tends to get easier for people to handle prepayment risk.

That is a key component of the equation.

Okay, that makes a lot of sense.

That helps.

Now have you guys looked at Ginnie Mae HECM pool?

Does it make any sense given the lack of sensitivity, rate changes with respect to prepayments?

Or does that not fit in to your overall investment portfolio thesis?

We have looked at them periodically.

I think we have focused our attention in other areas.

Generally speaking, those are interesting and I definitely could see why people see value there.

From our perspective, the other side of the Ginnie Mae equation is there is, generally speaking, a pretty big premium for just having, we will call it "the absolute full faith and credit of the government" which is really the key reason why Ginnie Mae's are generally off the radar screen.

I guess another way to look at it, the premiums may diminish some of the IRR over time.

The last thing I had Gary, on the mark on the Swaptions book, does that run through OCI or net income?

It runs through net income and GAAP terms, so that has already essentially been taken.

The Swaptions book is marked at this point at $5 million.

And, look, going to your question about a big increase in interest rates, while clearly those are way out of the money at this point, this is exactly what we like about the Swaptions.

They could have provided us hundreds of millions of dollars of value if the prognostications of higher rates came true.

And the worst case was we could lose $30 million or whatever, $0.15 a share.

And so big picture the Swaptions are essentially marked close to zero and if we had some disaster scenario they would actually kick in and they still have some value.

All right, thanks very much.

Thank you.

We have time for one last question.

Your last question comes from the line of Dean Choksi with UBS.

Thanks for keeping the call this long and letting me in the queue.

I have a couple of questions on prepayments.

Looking at slide 7, I guess the chart on the left, you have prepay speeds for 2010 in the 4%s.

You are really seeing that prepays for the more generic collateral really accelerated over the last couple of months, particularly in October.

I guess the first part is do you think the October prepayments for generics were an anomaly or the new trend?

And second, if it is the new trend, at what kind of mortgage rate do you think low loan balance, what mortgage rate would it take for low loan balance CPRs to really increase?

And then I guess the second question would be on the 13% CPR assumption you are using in your model, and that is a lifetime assumption.

So can you provide a little more color around how you think estimated prepays will trend around that lifetime assumption?

Yes, sure.

So that is a couple of questions so let me start with the first one which related to was October a one-time issue or where do we think things go from here.

So for that one the short answer is no, we don't think it was.

There will be noise.

There are a lot of factors that affect each individual month's prepayment, things like day count is a really important one, right?

In some months there are X number, there are two or three business days more or less than other months.

That can be a big factor, whether holidays.

There can be rate moves from a month or two earlier.

All of those effect any one particular month's print, so to speak.

On the other hand, we believe that you are still going to see higher credit, newer mortgages pay fast, for months, or until the interest rate environment changes.

And so we don't believe that is an aberration.

We actually think the trend is most likely higher.

Now let's take that to the next part of your question which is a very important one which is what happens with low loan balance.

And the short answer is we think there are two effects with respect to low loan balance.

The first one is that there are transactional costs that because these are smaller loans that make it, that they need a much bigger incentive, usually 50 or 75, it depends on the size of the loan, or more than that in terms of a rate incentive before they kick in.

So you have a pretty good cushion there.

The other effect is that there are just different types of borrowers and they just don't tend to get nearly as fast.

And then lastly there is a key component that helps the speeds on low loan balance which is a little different and we have mentioned this is in the past, but it becomes most important in environments like this where prepayments pick up.

And that is the issue of the capacity constraints.

And when an originator, a Wells Fargo or a Chase, or a mortgage broker, the guy calling people, or a third party originator, when these guys are out there trying to get business, where they have more business than they can handle, they are just not going to spend any time on a small loan.

They are not going to even, like the guy with the $100,000 loan gets put on hold for six hours while the $400,000 loans get processed.

It is just simple math and simple economics.

People get paid in that process based on the percentage of the loan amount.

And people that are busy are not going to process multiple loans to make less money.

And so that is sort of kind of one thing that really helps with respect to this [front] in that almost the worse things get on prepayments, the more important that factor becomes.

So keep that point in mind as well.

Thank you.

And then I guess how you expect the prepays to trend around that 13% lifetime CPR?

Gary Kain; Oh, thank you.

I'm sorry.

Our lifetime CPR, and this is going to depend on the product and how all of the mortgages and what type it is and so I can go on with all of the caveats.

Basically it allows for noticeably higher CPRs upfront than the 13.

It expects spikes that are, we will say materially above that.

And that assumes things burn out and start to come down from there.

And again, that is for most products and so forth.

The other thing to keep is it is informed by sort of the forward curve so there is also a built-in assumption on the backend that rates are higher and that prepayments are slowing for that reason.

So you can think of that, and this is an important point, that that assumption has already built into it speeds we will say over the next couple of quarters well in excess of 13.

So I think that is the best I can do in terms of giving you kind of a time profile related to that.

But I think that should help.

I appreciate that.

Thank you.

Enjoy your lunch.

I will.

Thank you.

And I appreciate the questions.

Thanks Terrence.

If you can close the call that would be great.

This concludes today's AGNC Shareholder's Q3 2011 Conference Call.

You may now disconnect.