Remember inverse floaters? In the 1990s they were the focus of hysteria on the part of some journalists and regulators who deemed them to be toxic waste investments. Recently some have used the word "toxic" when referring to mREITs. Is history about to repeat itself? For those who may not remember, in the 1990s inverse floaters became very popular with bond mutual funds. When interest rates rose, many of the bond funds lost value, faced large redemptions and in many cases were forced to sell their inverse floaters at distress prices. Litigation ensued.
Agency inverse floaters were issued by Federal National Mortgage Association Fannie Mae (OTCQB:FNMA) and Federal Home Loan Mortgage Corp. (OTCQB:FMCC) by carving out tranches from the large pools of mortgages they were securitizing. A typical inverse floater would be formed by taking a portion of the pool and creating an inverse floater and a companion floating rate tranche. One could take $9 million of thirty-year mortgages with an 8.0% coupon and issue $8 million of floating rate pass-through securities that paid LIBOR plus 1.4% and $1 million inverse floaters that paid 60.8% - 8 X LIBOR. Thus, the interest paid on the underlying 8.0% securities was being divided between the floaters and the inverse floaters. These were the terms of two tranches of a large issue (FHG 33) by the Federal Home Loan Mortgage Corp. (OTCQB:FMCC) in 1994. On the day that it was issued, LIBOR was 3.75% thus, the floaters paid 5.15% and the inverse floater paid 30.8%.
Note that the 30.8% paid by the inverse floater was exactly what could be obtained by buying $9 million of mortgage securities with an 8% coupon and using leverage to finance the purchase of the $9 million of mortgage securities by borrowing $8 million at a financing cost of 5.15%. Sounds familiar?
That is exactly how agency mREITs achieve their high yields. They leverage agency mortgage securities and finance them at low short-term rates. Thus, in the terminology of today's mREITs, the inverse floater FHG33s described above was achieving a 30.8% yield by financing agency mortgage-backed securities with a spread of 2.85% between the 8% coupon and the 5.15% borrowing cost and using $8 dollars of borrowing for every $1 of equity.
The 1990s inverse floater experience resulted in a number of mutual funds being sued and closed. Time magazine had an ominous story, "The Devil's in the Derivatives" on October 10, 1994, demonizing inverse floaters. What some may not be aware of is that every single agency inverse floater, including those that traded as low as 10% of face value at the height of the witch-hunt hysteria, paid 100% of face value within a few years after the litigation and bad publicity ended. Those who bought what regulators and others were calling toxic waste ended up with huge profits.
What all of this suggests for mREITS today is that there is the possibility of more pain, especially if a witch-hunt type hysteria develops. However, there is also the prospect of huge gains from the mREITs if short-term rates remain low for an extended period. In the 1990s, short-term rates did rise significantly and reached the point at which the coupons on many inverse floaters fell to zero as per their formulas. Had short-term rates not risen in the 1990s, holders of inverse floaters would have never have had any problems.
A number of years ago someone asked me for my opinion of Annaly Capital (NLY), which I had never heard of at the time. After a cursory look, I replied that NLY was essentially like an inverse floater. That was a somewhat dismissive statement as it was understood in that conversation that describing NLY as an inverse floater implied that there was considerable risk if short-term rates were to rise.
A few years later I became convinced that short-term rates were likely to remain low for an extended period (see my article: Federal Reserve Actually Propping Up Interest Rates: What This Means For mREITs), and I concluded that agency inverse floaters would be the ideal investment vehicle to take advantage of that scenario. I was told by dealers in agency securities that there were very few agency inverse floaters around anymore and that you would not want the ones that were.
Later, I came to regret that I had forgotten about the conversation where I had disparaged NLY as an inverse floater. I eventually did buy Cypress Sharpridge Investments (CYS), American Capital Agency Corp. (AGNC), Armour Residential REIT, Inc. (ARR) and ETRACS Monthly Pay 2xLeveraged Mortgage REIT ETN (MORL), but much later than if I had recalled my conversation about NLY being like an inverse floater and had connected the dots earlier.
There are similarities and differences between agency mREITs and inverse floaters. Both use leverage to obtain high yields. However, the distinctions between them can be important. Regulators and other critics of inverse floaters in the 1990s were confused by the fact that an inverse floater would generate exactly the same income as a leveraged account consisting of the underlying securities financed by a margin loan. Using the above mentioned FHG33s as an example, a margin account that contained $9 of a 30-year mortgage security with an 8.0% yield financed with $1 of equity and $8 of margin loan at a rate of 5.15% would yield 30.8% on the $1 of equity. That is exactly what the inverse floater would generate.
The regulators and others incorrectly concluded that the inverse floater had the same risk as the leveraged account with respect to the fact that if the underlying 8.0% 30-year mortgage security were to lose 11% of its value, the margin account would be wiped out, since the equity would fall to zero. However, no matter how much the underlying 8.0% 30-year agency mortgage security were to lose in market value due to rising interest rates, the inverse floater would still have value since it would still pay 100% of face value whenever the mortgages in the pool eventually matured or prepaid.
The mistake that regulators and others made was incorrectly thinking that the inverse floater consisted of only two components: the underlying mortgage securities and the loan element created by the issuance of the companion floating rate tranche. However, an inverse floater actually has three elements. In addition to the underlying mortgage securities and the loan element created by the issuance of the companion floating rate tranche, there are imbedded puts on LIBOR.
The imbedded put on LIBOR arises because the loan element created by the issuance of the companion floating rate tranche differs from a margin loan because the interest rate on the loan element created by the issuance of the companion floating rate tranche, is capped. Furthermore, the principal on the floating rate tranche loan is only repaid when the underlying mortgages are paid off.
If LIBOR were to go above 7.6%, the coupon on the FHG33 inverse floater be zero and the rate on the floater would stay at 9%, as all of the interest from the underlying 8.0% mortgage securities would be diverted to the floating rate tranche and none went to the inverse floater. Thus, there is a put on LIBOR with a strike of 7.6% embedded in the inverse floater, which protects the holder of the inverse floater from any increase in LIBOR beyond 7.6%.
The loan element created by the issuance of the companion floating rate tranche has its interest rate capped and has a term that lasts as long as the inverse floater. Thus, as opposed to borrowing via a margin loan or repo financing, there is no refinancing risk associated with agency inverse floaters. Essentially the buyers of the companion floating rate tranche have made a long-term loan to the holders of the inverse floaters to finance the purchase of the underlying mortgage securities. That loan floats with LIBOR up to a cap and the principle on that loan is only repaid as the underlying mortgages pay back their principle either from prepayments or the regular principle payments on the mortgages.
In terms of whether a witch-hunt type hysteria could occur today with mREITs, a big difference is that mREITS do not face the problem of shareholder redemptions that the open-end bond mutual funds did in the 1990s. A leveraged position consisting of mortgage securities financed by short-term repo financing, which is what an agency mREIT is, does face refinancing risk. The agency inverse floaters did not have refinancing risk. The holders of the companion floating rate tranche were essentially making a loan to the inverse floater holders for the life of the underlying mortgages.
The holders of the floating rate tranche were paid a rate above LIBOR for locking up their money. The above mentioned FHG33 floating rate note paid LIBOR plus 1.4%. Today the short-term rate on repos that mREITS use are very close to LIBOR. Thus, a mREIT willing to take all of the risks that are inherent with continually rolling over short-term financing could yield even more than an inverse floater based on the same underlying securities. However, the managers of mREITS use and pay for various hedging techniques to essentially make the mREIT more like an inverse floater in terms of refinancing risk.
Some mREITS try to replicate the embedded puts on LIBOR inherent in inverse floaters by buying puts on or selling Eurodollar or Federal Funds futures. Swaps and swaptions also hedge against the risk of short-term rates rising. These hedges cost money and mREITS disclose both the interest rate paid on their short-term borrowing, which has been running near .4% and their total effective borrowing costs, which includes both the interest rate paid on their short-term borrowing plus the costs of hedges, which attempt to lock-in some of the short-term rates.
While hedging can mitigate the impact to the earnings of an mREIT of an increase in interest rates, it cannot eliminate the refinancing risk that the mREIT could not rollover its short-term repo loans at any price. An interesting question is why mREITS do not use inverse floaters instead of trying to replicate the aspect of inverse floaters that limits refinancing risk. One potential advantage that mREITS have over inverse floaters is that if long-term rates increase, an mREIT can purchase higher yielding mortgage securities and actually increase the spread between the interest rate it receives and what it pays. In theory, the managers of an mREIT could achieve the limitation on interest rate and refinancing risk inherent in the inverse floater structure in a more cost effective way than inverse floaters do.
The bottom line is that a witch-hunt type hysteria could occur today with mREITs. The rout in mREITs due to recent fear that the Federal Reserve may be raising rates sooner than later might attract the attention of regulators and class action attorneys. The ensuing bad publicity could exert further downward pressure on mREIT prices. As with inverse floaters in the 1990s, such bad publicity could represent a tremendous buying opportunity.
If one believes that short-term interest rates will remain low for an extended period, mREITs represent a good way to obtain extraordinarily high yields. Although if there were actual agency inverse floaters available now, they would be my first choice to invest in. In addition to the risks of higher interest rates, there are always some firm specific risks with mREITS that would not be present with an actual agency inverse floater. The possibility exists that the management of an individual mREIT could make a major mistake in its hedging strategy or just a plain old fashioned bad bet on the markets. As I explained in 30% Yielding MORL, MORT And The mREITS: A Real World Application And Test Of Modern Portfolio Theory, the most efficient way to invest in mREITS in terms of risk and return would be to use baskets of mREITs such as Market Vectors Mortgage REIT ETF (MORT) and ETRACS Monthly Pay 2xLeveraged Mortgage REIT ETN (MORL), which are market-weighted baskets of 25 mREITS.
A risk averse investor could have some of their portfolio in MORT and some in a risk-free asset like t-bills. A more aggressive investor could buy MORL, which is leveraged 2X and generates more net income than most investors can generate from buying mREITs on margin since MORL borrows from its sponsor UBS AG (UBS) at .40%.