American Capital Agency Corp. (AGNC) slashed its dividend by 24c from $1.05 per quarter to 80c. For those who read our previous article, we surmised that the steady state, or sustainable, dividend for AGNC was 75c/shr and that at $3/annum and a 12% yield, we thought a $25 share price was right (more on that later). Well, that's as far as we'll "gloat" because, truth be told, we thought that AGNC would glide-path to that dividend level and were thinking something more like 90-95c for Q3. Even though taxable income and earnings wouldn't support the Q3 dividend fully, we thought they would return some capital and/or previously earned income to shareholders to smooth out the process. So, was it just the case of not suffering through 1000 cuts with 5c dividend reductions each quarter? Or should we assume the worse and run for the hills? Our guess is that, coupled with the announced 11.9mn share buyback at well below NAV, that AGNC is protecting longer-term shareholders while setting the dividend at a level commensurate with their current earnings power.
First, a few things to note. AGNC stated in their recent 10-q that as of June 30, 2013 they had distributed all of their 2012 taxable income and that they had an estimated $425 million of current year undistributed taxable income (UTI) , net of dividends declared. So, in other words, all of the UTI was and is available for paying distributions although they do not necessarily have to be distributed until AGNC's 2013 tax returns are finalized (late in 2014). So in effect, AGNC can distribute money from the kitty but, in general, one would want to keep a buffer until the results for the year are finally in. The other thing to note is that Gary Kain, AGNC's President and CEO, stated on their Q2 conference call that, "our taxable income in the third quarter is currently biased materially lower given some of the actions we took in both the second quarter and third quarters, which will be realized for taxable earnings purposes in Q3."
So as of June 30, AGNC had $425mn of UTI. We estimate that taxable income for the quarter was about $200mn. That is a 50+% drop from Q2 and in line with what they recorded in Q1 when they had some significant mark-to-market losses. Also, the $200+mn decline from Q2 would represent about 1/3 of the equity value decline witnessed over the course of Q2 when their hedging actions triggered taxable events (in this case losses but deferred until settlement in Q3). If we are right, that would give us about $625mn or so of UTI at the end of Q3 and, with an 80c dividend on ~385mn shares after the stated buyback, would result in about $317mn left in the kitty or about 82c/shr in UTI for the balance of the year. So even though the taxable income for Q3 would be only 52c/shr, lower than the 80c dividend declared, they have more than enough earnings from the first half to cover the balance. Could they have paid more, as in the full $1.05 original dividend? The answer in our estimation is yes. But that would have left only $200mn in UTI for the remainder of the year; not prudent from a risk-management perspective.
So why didn't AGNC use some return of capital to marginally decrease the dividend instead of this more severe cut, especially since they receive a huge amount of principal paydowns each quarter (e.g., $5bn+ in 1H13)? Well they could have. The incremental $100mn that would have been paid out to shareholders, assuming no cut in the dividend, is pretty small potatoes in the scheme of things. Instead, we believe, AGNC took a longer-term perspective. First, they decided to set the dividend at a proper level for the current book of business. That way, sans a repeat of the bond market debacle of the last few months, they could potentially surprise to the upside. Second, via the buybacks, they took advantage of a lower cost way to reinvest proceeds from their portfolio. AGNC bought back 11.9mn shares at an average share price of $22.16 which was a 13% discount to their stated NAV. Given the choice between reinvesting the $263mn in MBS at market prices, AGNC took the step of effectively buying more of their own portfolio at a discount of about 2%. That is based on data as of June 30 and since we don't know where MBS were trading at that time, nor their NAV, suffice it to say that is an estimate.
Of course, this then begs the question: why didn't they buy back more since it was such a bargain? Well, they could have but it increases leverage, all else equal, so there reaches a point where buybacks can be counter-productive. That said, looking at AGNC's leverage ratios below, we can see that the investment portfolio to equity ratio (row labeled "Portf Lvg"), while slightly higher than before the buybacks, is still well below what it was at its peak at the end of 2011, and at the end of 2012. We assumed that proceeds used for the buybacks came completely from MBS paydowns, as opposed to from cash say, and that all else was the same as of the end of Q2. Given where mortgage rates and spreads are today, we do not expect significant market related NAV changes this quarter. So it appears to us that AGNC made the reasonable assessment that, to benefit longer term shareholders, reinvesting a small percentage of MBS paydowns via buying back shares will prove beneficial.
AGNC Assets, Portfolio, and Shareholder equity ($bn)
So all told the actions taken by AGNC management gives us some comfort that we are on the right track and we do not expect meaningful surprises going forward. Given the implicit interest rate risk in mREITs, we feel that a 12% yield is acceptable although a discount for solvency risk would be reasonable. Given AGNC is trading in the high 13s on yield and a 5-10% discount to NAV, it seems priced with more than enough risk-premium. In light of the Fed's no-taper decision, it seems that rates, while gradually increasing, should not have another bout of bond market pneumonia. In fact, we would argue that the reason the Fed did not taper is because bond market expectations had gotten well ahead of the Fed's own forecast so we would expect the Fed to continue to govern any bond market sell-off that prices in "too much." But we'll leave that to another article. For now, we are happy with our initial position in AGNC and will look to potentially add on material weakness.