I was asked by one of my followers the other day to take a look at the future prospects of American Capital Agency Corp (AGNC), which is one of the larger mortgage-based REITs.
The question was posed as follows: Given the assumption that things keep going the way they are, what will be the price of AGNC in the early 2014 time frame?
Keeping in mind the well known phrase, that past performance is no indicator of future results, we can make an educated attempt at the project by taking data from the last 5 quarterly reports, which are posted here.
|Selected Financial Data for the last 5 Quarters|
|Shares Outstanding M||90||130||180||210||241|
|Portfolio Value $B||28.192||39.9||41.9||54.8||80.6|
|Interest Spread %||2.58||2.46||2.14||1.9||2.31|
|Interest Income M||164||201||232||263||406|
|Dividend Paid $/share||1.4||1.4||1.4||1.4||1.25|
|Book Value per Share||25.96||29.06||27.71||26.9||26.76|
|Stock Price $/share||29.2||29.74||29.71||28.06||30.07|
|Comprehensive Income M||154||244||273||476||587|
AGNC has adopted a growth strategy: The company has done a new share issuance roughly every quarter for the past year. In fact, just today, the company has announced the issuance of additional shares to raise capital for growth.
Over the past year, the company's number of shares outstanding has gone from 90 to 241 million, and their portfolio of mortgages has gone from $28B to $80B.
This is useful information because it gives us some measurement of the incremental money-earning portfolio value that AGNC can get for any future issuances.
I've done a lot of calculations, and made a lot of assumptions, and I will not bog the readers down at this point on the nuances of all of this, but I have posted on my instablog a lot of this information for inspection for those who are interested.
I will also be delighted to send any interested readers a spreadsheet with the key numbers so as to have someone comment and/or double check my figures.
The key to the calculation is that AGNC has been yielding upwards of 18% for the last year, and knowing what the dividend payout per share is, we can work back to some estimated stock price.
Naturally this is not 100% an indicator, as the students of the Dividend-Growth model know, but it is a starting point.
Scenario 1: Base Case
The assumption is: things keep going like they are. The interest rate spread remains the same (keeping in mind that the hedging results for AGNC are included in this number) and the marketplace stock price to reach an 18% yield, which remains constant throughout the period.
I've done two calculations: The future price, assuming that AGNC continues its strategy of share issuances each quarter, and the price with no share issuances.
There is actually price decay in this scenario and the issuance of the new shares actually depresses the price even more. The reason: the cost of the 18% dividend for the incremental shares is not paid back by a spread that is so low.
No one should feel bad about what is going to happen to the shareholders though: I am actually seeing a chance of a short term dividend increase. AGNC was a rock-solid 1.40 per share dividend stock for a couple of years before the first quarter of this year. The 1.25 dividend was continued in the second quarter. The model is suggesting that if things went equally well with the AGNC hedging program, there may be an increased dividend. The problem becomes whether the dividend stream plus the price decay equals a positive investment over the time period.
2. Decreasing Interest Spread
This scenario should look familiar: It is exactly what happened during the second half of 2011. Interest rates went down, and AGNC continued to do the share issuances to try to outgrow the problem.
The green line, above, should look very familiar to the shareholders of Annaly Capital Management (NLY) because the no-share issuance slow growth strategy is exactly what they used.
In this case the model still predicts a couple of good quarters of dividends but is not looking good at the end of the period.
3. Increasing Interest Rate Spread
The news is not always bad. With interest rates at their all time lows, it is possible for the market to deliver a surprise on the upside. In addition, since the hedging activities of these companies is in the interest rate spread number, this could possibly represent a good 18 months of successful hedging activity.
(click to enlarge)
The obvious question is: What is the likelihood of this happening? This is a matter of trying to predict a lot of externalities, such as government policy and events in far flung parts of the world.
4. Increasing Leverage
Here is how it works: The fund managers accept more risk, and are able to expand the portfolio more for the amount of money they take in. If you look in the table above, you can see that this is also what happened in the first quarter. AGNC's leverage increased to 8.2:1, so the company is already nibbling at this. The scenario below is what happens if they go all the way up to 9:1 leverage:
(click to enlarge)
This is a pretty successful strategy in keeping the price of the fund stable as long as the conditions remain stable. Here is the dividend stream:
If I thought the company was going to do this I would put 100% of my 401K into it because it does give a stable stream of nice dividends. However, there are no guarantees on anything, and by accepting more risk there may be some other underlying effects that are not predictable by this model, such as a 2008-like leverage-driven price collapse. Beware.
5. Yield Decay
In this scenario, the fund yield drops from an average of 18% which is how it has been running historically, down to about 16%. Before you say this is bad news, consider that a really good way for this to happen is the fund price going up, reducing the effective yield of the dividend. The proverbial price/book ratio that is talked about a lot is another artifact of a generally rising price. In fact, this scenario is exactly what is happening with AGNC right now:
I actually like this one a lot, because it sets up a bit of a positive feedback loop. The fund price goes up, each successive issuance raises relatively more money, which can be used to expand the portfolio even more. Naturally this, in combination with the rising spread scenario is looking especially good.
This scenario is just as good as it applies to the dividends.
The drawback to this scenario is that it is dependent on the market, which is fickle. Over the last several months, the market has seen fit to treat AGNC differently than it did a year ago, by giving it a favorable price. The relatively higher market price, even in the face of the dividend cut that happened in the first quarter, makes the effective cost of money lower and allows AGNC to get itself into this situation.
6. Preferred Stock
I didn't do the graph on this one, but I would just throw out for consideration that the issuance of $167M in preferred stock that the company did in April is an attempt to raise money at lower cost (only 8% instead of 16-18%). Thus far, this program is small relative to the rest of what is going on, but do not be surprised if the management tries to expand it.
So, what are we to make of all of this?
First of all, the strategy of issuing more shares every quarter is not foolproof, and in each of these cases above, given the leverage and rate of return that AGNC has, it does not make sense in all of the strategies. It may also not make sense for every fund because the tolerance for risk is a key factor, and that tolerance differs depending on the management philosophy.
In answer to the original question, as to what the future price of AGNC is going to be in the early 2014 time frame, the answer is "it depends". The most attractive scenario might be a case where the price is not too much higher than it is right now, but the nice stable stream of high dividends is paid. We are in an unprecedented era in history in which holders of capital are not rewarded for their deferred consumption.
Thirdly, in case you missed it above, I think AGNC is going to have a couple of pretty good quarters, and may even increase the already-nice dividend. Maybe the market is anticipating this. We will get more information when AGNC issues its quarterly report in a few weeks.
Fourthly, the farther out in the future you get, the less reliable this, or any other model might be, so take that into consideration before making any investment decisions.
Maybe I can make my little computer model generic, and try to forecast the performance of some of the other mREITs.
As we are so fond of saying, the world is chaotic, and there are no guarantees on anything.
Additional disclosure: I am also long ARR, CYS, IVR, MTGE and TWO.