As per the framework in the prior article, MBS spread should be the bigger driver of book value and dividend sustainability from here on under the assumption that management has stuck to keeping duration close to 0.
We can try and estimate what book value for Q3 2013 could be by adjusting the Q2 2013 book value by estimating for the following relevant changes:
- Changes in interest rates
- Changes in MBS spreads
- Adjustment for dividend paid
- Adjustment for share buy back
- Spread Earned
But before, it will be helpful to see what has changed during Q3 which could materially impact the calculation process.
What has happened in Q3?
- AGNC cut their dividend to $0.80 per share closer to the $0.60 - $0.66 I had estimated previously.
- More importantly, AGNC announced that in the third quarter, the company repurchased approximately 11.9 million shares for $263 million.
- US Treasuries have been volatile intra quarter but closed Q3 relatively unchanged.
Estimating Changes to Book Value
- Impact on Book Value from changes in Interest Rates on Net Duration.
The Q2 2013 10Q provides a good interest rate sensitivity as shown in table 1:
Table 2 shows the changes in Us 10year swaps:
So with interest rates having changed by +6.8bp, the impact of rise in interest rates should be around -0.30% to book value, so around -($0.08) per share.
Above proxy indicates that the change in book value from changes in interest rates should be marginal, however this does make 1 important assumption that management did not choose to readjust duration as rates were moving steadily higher in the first half of Q3. This would have been equivalent to increasing the hedge at higher rates and as rates reversed lower, the extra hedge would incur losses not included in the above calculation.
2. Impact of changes in MBS spreads.
Since AGNC now has a mix of 15yr and 30yr MBS, it may be worthwhile looking at 15yr MBS vs 7yr US swap rates and 30yr MBS vs 10yr US swap rates. I have included 2 charts to give an idea of the history of the 2 spreads.
Note that MBS spreads are towards the lower end of the range since the beginning of 2012.
MBS spread levels around the relevant quarter dates are given in table 3 below: (data is weekly so does not fall exactly on quarter end dates).
The weighted average column is provided to reflect AGNC's holdings which is 58% in 15yr MBS and 42% in 30yr MBS.
Unfortunately there has been a big move from June 26th to July 4th and with Q2 2013 falling in between, accuracy will be lost but we can consider the following 2 scenarios as given in table 4:
In calculating the impact of changes in MBS spreads on book value, the approximation formula of Duration * notional * basis point should be a good indicator.
Tables 5 and 6 below provide the approximate change in book value under the 2 different scenarios.
Under the conservative scenario, we can expect book value to have increased by around $0.47 whilst the average scenario would be an increase of $2.52.
Putting Everything together, using Q2 2013 Book Value as starting point.
Including impact from changes in MBS spreads:
From the above framework, it seems to me that the book value of AGNC for Q3 2013 could be in the range of $25.66 - $27.70.
Given the current discount of 6% to the lower end of the book value estimate, I have increased my exposure to AGNC through short dated options in the hope that either the book value for Q3 2013 comes higher or at the very least market reduces the discount priced in.
For Q4 2013, it would be interesting to get an indication of whether management would still maintain a neutral duration bias going forward. If so MBS spreads would still be the bigger driver of book value with MBS spreads currently at the lower end of the range since 2012.
Additional disclosure: Long options on AGNC