There has been a lot of interest in mREITs lately. They are attractive to many income investors because they have high yield and seem to have low volatility, exactly what income investors are looking for. Others are attracted to mREITs because their correlation to the S&P 500 is relatively low.
Of course mREITs have their risks as well. Many have pointed out that the current interest rate situation, with its upward-sloping yield curve, very low short-term interest rates and low repayment speeds is an ideal environment for mREITs to earn money. Bulls counter this argument by pointing out that ideal conditions look set to continue for at least 12 months and that in any case mREITs have done well through the tough times as well.
Regardless, there are and always will be risks inherent in the mREIT business, as with any business. Some of the major ones are:
- a new credit crunch in which the mREITs are no longer able to roll over their repurchase agreements, or can only do so at higher rates, thus reducing the interest rate spread from which they earn money;
- an event where yields on mortgage debt suddenly rise, thus damaging the book value of their current asset book;
- the management team just screws up on a derivatives deal; and
- a regulatory change that diminishes their profitability.
I am not here to debate the riskiness or otherwise of mREITs. I am investigating a risk-management strategy for AGNC to see if it can work.
In this article I consider a strategy for hedging the stock of American Capital Agency Corp. (AGNC). AGNC's dividend yield is $5 (over 15%) as of this morning. The object is to, at the cost of giving away some portion of the yield, cap my losses in the event that disaster strikes.
I have considered 4 scenarios, all involving long-dated puts.
- I buy a January 2014 $35 put. I evaluate position and yield 12 months from now.
- I buy a January 2014 $27 put. I evaluate position and yield 12 months from mow.
- I buy a January 2014 $25 put. I evaluate position and yield 12 months from now.
- I buy a January 2013 $27 put. I evaluate position and yield at expiry.
I have investigated what the situation would look like for a range of prices from $24 to $36 for AGNC at the evaluation time. This represents a range from about -2 to +1 standard deviations for AGNC's annualized historic volatility of 10%. Theory says that in 12 months' time, there should be about a 68% chance the price falls between -1 and +1 standard deviations (approx. $28.80 -> 35.20) and about a 95% chance it will fall between -2 and +2 standard deviations (25.60 -> 38.40), so my range has very good coverage on the downside and I am not so interested in the upside. (We will see that it is a straight line.)
I have made the following assumptions about how the stock behaves as it goes down and up in price and in evaluating the results:
- The stock price drops for a reason, and that reason is mainly dividend droop. I assume that if the price has dropped to $24, then its dividend has also dropped to $0.60/quarter or less than half of what it is now (10% forward yield at the $24 price). Total dividend for the next 12 months is $3.75 under this scenario. On the upside, I assume the quarterly dividend climbs to $1.57 for a total over the next 4 quarters of $5.62.
- I have used linear interpolation for dividends, across quarters and across the price range.
- For the January 2013 case, of course, the last of the 4 quarterly dividends has not been paid yet so I have excluded it from the analysis.
- Dividends are not reinvested.
- For the January 2014 puts, I had to make a guess at what their worth would be in May, 2013. I have simply taken the prices on the Jan 2013 strike chain, which has exactly 12 months' less time value. To try and work out what an option would be worth if the price of the stock changes, I have looked at the option price the same number of strikes in or out of the money. So for example, in trying to guess what the price of a Jan 2014 $27 put will be worth 365 days from now, when the stock price has risen from $32 to $33 between now and then, I take the price of the January 2013 $26 put from Friday's option chain. This of course has a number of theoretical flaws - for example it assumes constant implied volatility and it also assumes everything is nice and linear when it is actually at best piecewise linear. However, I contend that the systemic error is less than the noise in the system, and the method remains useful.
- Options are bought at the ask, and sold at the bid. Bid/ask spreads are actually the worst enemy here and must be taken into account.
- The position is closed or rolled over at the evaluation time. This is probably not the best strategy.
The results are summarized in the chart below, which is the risk profile for buying the stock at $32 (price as of Friday 11th May), along with the put options, evaluated on 11 May 2013 for the Jan 2014 options and on 19 Jan 2013 (expiry) for the Jan 2013 put.
Some observations about this:
- The January 2014 $27 put looks like the best hedging instrument to me. It actually caps loss at about -6.5-7%, while approximately halving the yield at unchanged price from 15% to 7%. Since delta is negligible on the option by an underlier price of about $33 and above, you get nearly all the upside after paying the premium.
- The January 2014 $25 put is only slightly better on the upside, but caps the downside loss to about -10%.
- The January 2014 $35 put performs badly right the way through, because an in-the-money put is just too expensive.
- The January 2013 $27 put looks a lot like the January 2014 $25 put, except that it hits maximum loss quite a lot earlier. You also have to consider that the analysis for this particular option is over 253 days instead of 365 days so the annualized risk profile will actually be somewhat steeper. (But then the price fall or rise also needs to be steeper to hit the price in a shorter time). I only included this so you could see the general effect of using shorter-dated options.
One of the biggest components of loss in this exercise is the bid/ask spread. Long-dated options for this stock are not very liquid and the bid/ask is huge. In practice, you can improve on this scenario somewhat:
- Don't roll over the option in May 2013. Wait until November when the bid/ask spreads for the option sale are narrower. Or just let the option expire and assign the stock in Jan 2014 if it is ITM or buy another hedge if you feel it is a good idea if it expires OTM. Essentially you get another 200+ days of hedge for your money.
- Do buy with limit orders. You will probably be able to get the put option cheaper if you are patient.
- If you buy 200 or more shares, then you can change the hedging ratio by buying fewer contracts and increase your yield if all goes well.
I also looked at doing this with NLY and Capstead Mortgage Corp. (CMO). It doesn't work very well with these stocks because the bid/ask spread is too large as a percentage of the stock price. If you could buy at the midpoint and hold to expiry, it may look better.
So is this a worthwhile strategy? Only you can be the judge of that. At the very least it is interesting. And if you buy AGNC with a January 2014 $27 put option, you get an instrument which yields 7% if the price does not change and has loss capped at about 6-7% even if disaster strikes. That is better and safer already than a 30-year bond, and the upside is much better. In these days when many are focused on simple capital preservation, it is certainly worth a look.