Please see my earlier post, on the creation of a Mortgage bond
which reviewed 3 parts of the mortgage bond market.
- “When Issued” forward-trading of mortgage origination supply
- The packaging of homeowner loans into plain vanilla mortgage bonds
- The role of mortgage servicers and mortgage insurers in the bond market
This post will cover subsequent features of mortgage bond trading and structuring
- The basics of mortgage trading – Prepayment risk!
- The basics of CMOs – mortgage derivative structuring.
- Recent market moves must have caused a bloodbath on most Wall Street mortgage desks
Mortgage Trading – All about the prepayment risk
So how do mortgage bonds trade on Wall Street? How do investors think about the product?
Contrary to popular reports, plain-vanilla mortgage bond trading and investing remains among the safest type of investing from a ‘credit’ perspective. Investors can always expect their full principal and interest returned to them.[1]
Our simple FN 8720331 4% bond issued in October 2013 offers investors AAA-risk comparable to US Government bond risk and extreme liquidity, meaning an investor can sell the bond at any time and not pay much in transaction costs.
The main and only significant difference between our FN 4% bond and a similar US Treasury bond is the uncertainty of the timing of principal payments. Meaning, the US Treasury does not generally pay back its bond principal early, but a mortgage bond, by contrast, pays a little bit of principal, every single month. In addition, if any of the 2,000 underlying homeowners decides to sell or refinance his house, an unexpected principal repayment flows through to the bondholder.
Not if, but when
The risk to the bond-holder, therefore, is not if he’ll get paid back, but when. Since in our example 2,000 individual homeowners have the choice over when, anytime in the next 30 years, their individual mortgage gets paid back, the mortgage bond holder is subject to other people’s choices, which the investor cannot control.
The mortgage bond holder, by purchasing the bond, has implicitly sold 2,000 little 30-year options to homeowners.
In financial terms, when you sell an option you get paid a premium for that option. That premium shows up in the form of extra ‘yield,’ or investment return over comparable US Treasury bonds. This makes plain vanilla mortgage bonds ‘yield’ more than other AAA-rated bonds, but it also burdens the bond holder with ‘prepayment risk.’ Mortgage investors and mortgage traders spend all of their waking hours stressing about prepayment risk.
NSA-sized computer servers, and the greatest minds of our generation are dedicated right now, as I write this, to modeling prepayment risk in mortgage bonds. I didn’t say they were tapping the phones of those 2,000 homeowners to get a sense of when they will refinance, but I mean seriously, do you doubt it? There’s a lot of money at stake here, after all.
The problem of being a mortgage bond holder, just to dig a bit further into the prepayment problem, is that homeowners always exercise their option to refinance their home at the precisely wrong time, for bond investors. What do I mean by that?
Always on the losing end of volatile markets
I mean mortgage bond holders get paid early just when they don’t want to, and they don’t get paid back early when they would like to be paid early.
When rates drop strongly, for example, many more of our 2,000 homeowners will choose to refinance early to take advantage of the new interest rate savings. That uptick in refinancings will send early payments to the bond holder and amortize his 4% bond more quickly than expected. Unfortunately for the dedicated mortgage bond investor, however, he has to own mortgage bonds, that’s what he does for a living. So he needs to invest in new mortgage bonds to keep his money earning money. He will have to reinvest his cash at the new lower rate, which might only earn him 3%, since rates dropped strongly.
Rate hikes can be even more deadly. If mortgage rates jump to 6%, for example, many fewer of our 2,000 homeowners than usual will opt for refinancing. In fact, only people who move houses (or get foreclosed on) will pay off their mortgage early.
Few homeowners elect to refinance into higher interest rate mortgages. Hardly any prepayments flow to the bondholders to amortize the bond. Our bondholder anticipated a certain amount of prepayments and hoped to invest his proceeds at the new higher rates. He can’t. In the new 6% interest rate world, new mortgage bonds pay close to 6% but our mortgage bond investor still holds our dumb old 4% bond, with less-than-anticipated cash to put into the higher yielding bonds.[2]
The more that interest rates move, the worse off the mortgage bond holder fares, in both directions, because he’s short the prepayment option.
In times of volatile interest rate moves, the homeowner holds this very valuable option, and the bond holder suffers as a result.
Everything about mortgage trading and investing flows from this fundamental problem – the problem of prepayment risk.
Mortgage Servicing and Mortgage Derivatives – attempts to solve prepayment risk
The 2,000 individual homeowners paying their monthly mortgages underneath our theoretical $750 million Fannie 4% bond are really raw financial clay with which Wall Street artistes create financial sculptures.
If prepayment risk is the ultimate heavy burden of mortgages, the point of the financial sculpture of mortgage derivatives is to shift risk in ways to defy gravity, ultimately matching investor appetite for prepayment risk.
The mortgage servicer who separately pays interest and principal payments to bond-holders plays a key role in making these works of art possible
Simple mortgage derivatives
CMO – A Collateralized Mortgage Obligation is a generic term for relatively simple mortgage derivatives, first created 30 years ago, that typically shift prepayment risk forward or backward in time over the life of a mortgage bond.
A Wall Street bank may decide to sculpt our FNMA 4% bond into a CMO structure to split the timing of mortgage prepayments.
As a simple example, let’s assume three different investors want three different types of investments.
What a savings and loan bank wants
A traditional savings and loan bank might be looking for a place to park its cash for up to 2 years and is happy to earn a safe 2% return on its money. Our theoretical bank investor needs everything it is investing in a mortgage bond returned over the next 2 years to make its budget, and it cannot risk tying up its capital much past the next 2 years. Our savings and loan bank needs a CMO structured to receive lots of mortgage prepayments.
What an insurance company wants
An insurance company, by contrast, typically seeks long-term bond investments to match its need to meet its long-term liabilities, like life insurance payouts. The insurance company seeks a way to invest its capital for 10 years, but needs something more than a bank for that long term investment – it seeks a 4.5% return. In addition, the insurance company really does not want to receive early principal payments. The point is to keep its capital earning the 4.5% rate for as long as possible, so the insurance company really wants a CMO structured to help it avoid prepayments.
What a hedge fund wants[3]
Finally, a hedge fund has a flexible view of yield and the timing of return of capital, but thinks it has a better sense than the rest of the market on the true likely prepayment speed of this FNMA 4% October 2013 cohort. The hedge fund wants to earn extra yield and is willing to stomach the risk of a wider range of bond payment timing outcomes. In financial lingo we’d say the hedge fund earns the extra premium by being “short” a volatile pre-payment option. By buying the CMO with the most volatile outcome, the hedge fund has done the financial equivalent of selling many call and put options to homeowners, and it hopes to profit from this implied sale, if the interest rate environment turns out to be less volatile than expected.
Our clever Wall Street firm can assign our FNMA 4% bond to a CMO structure and instruct the mortgage servicer to follow a set algorithm as prepayments arrive over the next 30 years.
All principal payments first go to pay the bank’s CMO until that bond is completely paid off, followed by the hedge fund CMO, followed finally by the insurance company’s CMO.
For the bank’s CMO all principal payments – both the scheduled principal amortization and the unscheduled prepayments – get forwarded to this short-term bank CMO. As a result, this bond pays down extremely quickly and will likely return all capital to its holder within the 2 year time frame. There is some uncertainty about timing, but the fact that the bank CMO gets every single principal payment really limits the prepayment timing to within a nice, tight, short range.
The bank’s CMO structure also makes the next two CMOs created from the same FNMA 4% bond possible.
The hedge fund CMO only receives principal prepayments after the bank CMO has been fully paid off. As a result, the hedge fund knows it will not be subject to prepayments for some period of time in the very near future. As it is second in line for principal payments, this CMO acts kind of like a shock absorber for the other two bonds, and will be quite sensitive to changes in interest rates and therefore prepayment speeds. The investor in this type of bond, like a hedge fund, will likely believe it has a better read on prepayment risk than others in the market. Because it takes on the most prepayment risk of the three bond structures, the hedge fund will demand the most yield enhancement over comparable AAA bonds to compensate for this increased risk and uncertainty.
The insurance company’s CMO, as third in line for prepayments, has two layers of prepayment ‘protection.’ Although the timing of principal payments may ultimately differ significantly from the insurance company’s expectations, the two layers of protection cushion the prepayment risk and keep it within a tighter range than would be otherwise available from a plain-vanilla 4% FNMA bond.
By slicing up our mortgage bond pool to meet the demand of three separate investors, the Wall Street firm can, ideally, sell the entire pool at a higher implied price than would be otherwise available in a plain vanilla format. Happy customers, and higher fees, follow.
Interest only bonds and principal only bonds – another simple CMO structure
Because interest and principal payments for our $750 Million 4% FNMA bond can be easily separated by the mortgage servicer, Wall Street desks quickly figured out that some investors want interest only bonds, while others prefer to receive only principal.
Who would want an interest-only bond?
The first feature of an interest-only bond is its potentially volatile and leveraged nature – it fluctuates widely in value if you get the bet right. The second feature is that it moves in the opposite direction of most bonds due to changes in interest rates.
Most bonds go down in value as interest rates rise. But interest-only bonds created by mortgage pools will increase in value as rates rise. That’s because we expect prepayments will drop with a rise in rates, which means that you will receive interest payments on the 2,000 underlying pools for a much longer time, as fewer homeowners extinguish their mortgage through refinancing.
The price of an interest only bond will shoot upwards if interest rates unexpectedly shoot upwards and prepayment expectations drop accordingly.
If I as a mortgage investor need to hedge my mortgage portfolio against an unexpected rise in rates, I might shop for interest only bonds. If my entire portfolio is likely to lose value when rates rise, I benefit from the hedge of owning bonds that rise in value when rates rise.
Conversely, a bet on a principal only mortgage bond may be a type of bet on a decline in interest rates. Principal only mortgage derivatives will be especially sensitive to changes in rates.
Principal only mortgages trade at a discount to face value. If prepayments arrive more quickly than expected (due to, say, an unexpected increase in refinancing activity) the principal-only mortgage holder wins. Every principal payment is made at ‘par,’ causing an investment gain versus the discounted price at which the investor bought the entire principal-only mortgage derivative.
If for example I bought my principal only bond at 80 cents on the dollar, but 10% of my 2,000 underlying mortgages prepay early this month, I’ll get 10% of my investment returned to me at 100 cents on the dollar. That’s a win.
Funkier structures – More CMOs
Traditionally, mortgage structuring desks attracted some of the brainiest folks on Wall Street.[4]
With the raw material of a home mortgages, the creativity of these artistes knows few limits. Some CMOs provide precise prepayment certainty to risk-averse investors, as long as other ‘companion bonds’ serve as shock absorbers for unpredictable prepayment risk. Companion bonds will be retained by Wall Street mortgage desks comfortable with the risk, or may be bought by hedge funds with a higher risk appetite or a strong conviction about future prepayment risk.
Some CMOs offer floating interest rate structures to investors seeking to eliminate interest-rate risk exposure, while creating ever-more algorithmically complex ‘companion bonds.’
For those curious about the ever-awesomer financial sculptures the smartest minds of our generation can create, I recommend this Wikepedia page.
Recent market moves must have been ugly
Interest rates shot up more than 0.5% in May and June. For mortgage bond holders, interest-rate volatility generally hurts, and rising rates provide a double whammy to the problem.
There is no doubt this kind of movement is career-making and career-ending for Wall Street folks. Rates have been so low, for so long, that some mortgage desks will be positioned right, and many more will prove in retrospect to have been positioned wrong.
If you held a preponderance of IOs, or some extraordinary floating rate structures, or got massively short interest rates in April 2013, you’re probably ok.
For a great number of mortgage investors and traders, however, I suspect they didn’t save themselves from huge losses.
Also, see previous posts on Mortgages:
Part I – I refinanced my mortgage and today I’m a Golden God
Part II – Should I pay my mortgage early?
Part III – Why are 15-year mortgages cheaper than 30-year mortgages?
Part IV – What are Mortgage Points? Are they good, bad or indifferent?
Part V – Is mortgage debt ‘good debt’ A dangerous drug? Or Both?
Part VI – Mortgage bond creation on Wall Street
Part VIII – The Cause of the 2008 Crisis
[1] Before you get smart-assed about all the scary bond losses you read about once in a Gretchen Morgenson article, let me reiterate that I said plain-vanilla mortgage bonds, not risky portions of mortgage-backed CDOs or sub-prime structured products.
[2] Adding insult to injury, if it’s an un-hedged mortgage bond position, his bond also trades significantly below par. So if he decides to sell the 4% bond to buy a 6%, he’ll take a loss. Rate hikes are hard on all bond investors, but especially mortgage bond investors.
[3] If you were born around the same time as me, it’s what you want Joel. Its what every white boy off the lake wants.
[4] Plenty of raw idiocy and barbaric types too, of course, like any testosterone-fueled environment, anywhere. For a humorous depiction of the brutal origins of the market, look no further than the original Michael Lewis classic, Liar’s Poker.
Post read (13145) times.