I’ve always been bearish on mortgage REITs. In March 2013, I discussed that AGNC was not earning its dividend sustainably and that a cut was likely. In June 2013, AGNC announced a dividend cut.
In September 2021, I explained why stock REITs are generally better investments than mortgage REITs.
An exception to my downtrend on the sector was an opportunity in preferred mREIT agencies such as ARMOR Residential (NYSE:ARR) preferred C and Annaly Capital (NYSE: NLY) Preferred D which became opportunistically cheap due to a price dislocation during the pandemic. Today, those preferred shares are all at or above their $25 face value or have been redeemed, so the opportunity has largely dried up.
I bring up this history to show that in writing a bullish thesis on AGNC today, it is not from a “yay high yield” perspective but rather a conscious analysis of the specific opportunity at this valuation. specific.
Given the general poor price performance among mREITs, some of them are now cheap enough to offer a decent total return. In particular, I move higher on AGNC Investment Corp. (NASDAQ:AGNC).
The purchase thesis
The market has sold AGNC so much that it is now almost the cheapest it has been in the past 10 years.
The sale seems to be related to 2 things:
- Dividend investor frustration with dividend cuts.
- Fear of rising interest rates.
None of these things have a significant impact on its fundamental value, so I’m assuming that at this reduced valuation it has a strong return going forward. The earnings yield is high due to the low market price and the sustainable earnings fully cover the 9.5% yield.
Cuts and fear of rising interest rates made AGNC cheap
The 2013 dividend cut was just the tip of the iceberg, with dividends falling precipitously.
With its monstrous dividend yield, AGNC was very popular among dividend investors. However, the chart above is the exact opposite of what most dividend-focused investors like to see.
As a result, many sold the name in search of more stable dividends.
Given the high level of return on AGNC, the total return was actually positive, but it still significantly underperformed the market.
The sale of AGNC accelerated in mid-2021 as it began to become clear that the Fed was going to have to take action to deal with inflation.
The basic idea here is that rising rates tend to hurt highly leveraged companies that use short-term debt.
All that selling pressure has left AGNC just about the cheapest in a decade.
It now trades at just under 6 times forward earnings and 87% of book value.
This level of cheapness is justified in the eyes of the market, as many anticipate further cuts in dividends and earnings due to higher interest rates.
I see things a little differently.
Historical dividend cuts are not a relevant factor in determining future dividends
As a REIT, AGNC is required to pay out at least 90% of its taxable income in the form of dividends. For equity REITs, this is not a problem because they have depreciation to shelter their cash flow, allowing the company to retain significant cash flow beyond their dividend payout. Mortgage REITs like AGNC don’t have significant amortization, so they can’t shelter their income and have to pay just about everything. There’s nothing wrong with shareholders receiving the bulk of profits in the form of dividends, but there’s a big problem when taxable profits far exceed actual cash profits.
This was the case for AGNC for much of the past decade and the reason it had to cut its dividend. AGNC therefore has 2 major income brackets.
- Net interest margin or NIM which is the difference between what they earn on their assets and the cost of capital
- Mark to market of securities
The first bucket is what I would call true profits while the second bucket is just a matter of market fluctuation.
As interest rates have fallen significantly over the past decade, the market price of RMBS securities has risen such that the yield to maturity would drop to market rates. This price increase caused AGNC to make huge gains in the market price range.
These gains forced AGNC to pay out a dividend far in excess of their true first bucket earnings. Since interest rates cannot fall forever and therefore positive market value cannot last forever, a dividend cut was inevitable.
This is why there has been a reduction and it is also why historical dividend cuts do not predict future cuts.
Today, AGNC’s dividend is covered by NIM’s actual earnings and therefore looks quite sustainable. In fact, NIM has improved considerably and is at a high level.
Interest rates – threat or not?
Based on AGNC’s market price action, it is clear that the market perceives rising interest rates as a threat.
However, looking at the mechanics, I see it as non-threatening.
Thus, AGNC’s net interest margin operates from their assets generating a return in excess of their cost of capital. Given the current zero interest rate environment, Repo rates are close to 0.
This is the bulk of AGNC’s cost of capital.
The fear of an interest rate hike is that when the Fed rises, repo rates will likely rise in line with the fed funds rate. I agree that it might happen.
However, with the rate hike telegraphed by the Fed for a long time, companies were able to prepare. Below is AGNC’s hedging activity.
Overall, AGNC’s cost of capital is 98% hedged. So even if interest rates rise, AGNC’s cost of capital is largely locked in.
The other impact of rising rates is that it could lead to a negative rating in the market. Just as falling rates shifted one-time income from a positive mark to the market, rising rates would cause one-time income to be negative. Here is the sensitivity table.
I find the full 75 basis point hike to be quite plausible and it would drive the book value down by 9.6%.
Well, with AGNC trading at less than 90% of book value, that success is already fully priced in.
The key thing to note here is that this hit would be one-time in nature and would not impact their dividend coverage. It’s coming from the NIM and it looks healthy whether rates are rising or not.
The real risk that investors should think about is an explosion in spreads.
Hedging activity handles parallel shifts very well when all the financial parts move properly in tandem, but it does not handle spread breakouts.
You will find below the table of sensitivity to spread variations between RMBS and Treasuries.
If, for any reason, market yields for RMBS rise significantly relative to Treasuries, it could result in substantial losses for AGNC.
Such movements are usually the result of a large increase in risk premia and I don’t think that is likely. Given the general economic stability and the telegraphed nature of the Fed hikes, I think the 30-year mortgage rate and Treasuries will rise in tandem.
So far, that has been the case this cycle.
In fact, I think the spread will tighten slightly as the RMBS is moderately de-risked.
With interest rates at 0, refinancing mortgages became very lucrative for homeowners, driving refinance rates up to 40%.
As interest rates rise, refinancing becomes less of a financial incentive and the pace of refinancing should come down. This movement has already begun.
With more moderate rollovers, the duration of the RMBS is more predictable and there will be fewer par premiums lost in the event of early redemption.
I expect the gains from the spread tightening between RMBS and Treasuries to more than offset the market price loss from higher interest rates.
The overall picture
Given the sell-off, AGNC is now trading at an attractive valuation. With a strong earnings yield and a steep discount to the market, I think this is a good entry point. A healthy net interest margin and abundant hedging activity ensure that the 9.5% dividend is fully covered by sustainable cash flow.
The best time to invest is often when sentiment is at its lowest, especially when sentiment is more bearish than actual fundamentals. Given the general disdain for AGNC from previous dividend cuts and fears of rising interest rates, I think the market is far too bearish on AGNC.
This has driven the price down to a point where total forward returns appear to be beating the market.