Hello class. I'm @tdgraff and will be your guest lecturer in the BBB series today on Mortgage Backed Securities. For background, I'm a buy-side PM today but I actually came up as an MBS analyst, and one of the strategies I run is mortgage specific.
I'll try to live up to @EffMktHype's standards, but no promises :)
We'll cover what MBS are, how they trade, how you analyze them, who the buyer base is, what a CMO is, when MBS tend to outperform, and whatever else the audience asked about.
@EffMktHype I'm going to focus on "agency" MBS, which are the ones backed by Fannie Mae, Freddie Mac or Ginnie Mae. Non-agency MBS (like what blew up in '08) are a far smaller portion of the market, and TBH should be covered in a discussion of ABS/CMBS, which I'm not going to get into here.
@EffMktHype First, the MBS market is huge, at about $11.5 trillion according to @SIFMA. That's about $1 trillion bigger than U.S. corporates, and only topped by Treasuries.
It is also probably the second most liquid market in the world, after U.S. Treasuries. Again according to SIFA, about $300 billion of MBS trade per day, vs. just $40 billion of corporates, $9 billion of munis. Again, only topped by Treasuries (~$620 billion).
Mortgages that are sold to investors are bundled into "pools." The pool might hold as few as a dozen loans or it might hold 100,000+. The bank that originates the loan will sell the loan to one of the GSEs, and will pay a "guarantee fee" before selling it to the public.
Once that happens some bank will become the "servicer." This could be the originating bank or someone else. The servicer is who actually deals with the borrower from here on. I.e., that's who you make your check out to, who would deal with you in delinquency, etc.
Basic MBS are often called "pass throughs." This is because the cash flow "passes through" from the actual borrower to investors. When you send in your monthly mortgage payment some of that is principal and some is interest, which we'll call "P&I" from here on.
The servicer takes a fee out of the interest, then the rest of the interest plus the principal "passes through" to investors. Whatever principal is paid actually reduces the amount of bonds you own. The percentage that is remaining is called the "factor" on a mortgage.
Quick aside: the factor bit makes MBS trading easier. When we transact, the quantity quoted is in "original face" which is how much was originally outstanding. Then the factor is applied.
Below is an example using the "BXT" function in BB.
In this example, I'd tell my counterparty that I'm buying 1mm of FN BQ8987 (which is the pool ID) at a price of $104.766. Since this has 98.100563% outstanding, I actually wind up with $981,005.63 "current face."
The GSE's role in all this is strictly as guarantor. If a loan goes delinquent, at first the GSE advances both P&I to investors. However if a loan goes more delinquent for 4 months, the GSE "buys" the loan from the pool at 100% of its original face.
So in essence, any defaults act just like principal repayments from the investors' perspective. Worth noting that long-term delinquencies are rare. According to Fannie, about 1.7% of loans from 2009-2021 are seriously delinquent.
You can see the credit stats for a given pool using the CLP page in BB. This pool currently has zero delinquencies:
I've written about this before, but for those new to following the yield curve, here's a simple way to think about the dynamics between long and short rates. I think this will show you why this bear steepener has an expiration date. 🧵
We start with a basic assumption: each yield along the Treasury curve reflects the market's assumption of the average Fed target rate over that period of time, plus some risk premium.
Why would this be? 2/
In a simplistic world, investors are choosing btw buying a series of overnight bills OR locking their money up. I.e., I can buy an overnight T-Bill every day for the next five years OR buy the 5yr Treasury.
My expected return on both has to be close or there's an arbitrage. 3/
For the 10yr TIPS to trade with a breakeven of 3.00, markets have to assume that the *average* inflation rate will be 3% over the next 10 years.
Is that possible?
I suppose in an infinite universe all things are *possible* but boy... seems quite *implausible*. 2/
There are two major reasons why TIPS would trade with ~3% breakeven. The first is that the Fed has basically given up on its inflation target of 2%. While the Fed's target is actually based on Core PCE inflation, the gap to Core CPI isn't going to be 100bps. 3/
If I were running comms at the Fed, I'd want Waller giving this speech. But since I'm an investor, I don't think there's any signal here. 1/ reuters.com/markets/us/fed…
First, Waller is throwing a bit of a strawman out there. "This isn’t ending in the next meeting or two." The market knows that. It has priced a 50bps hike in Dec and 2x 25bps in Feb and March. The market isn't "way out in front" in the way he's claiming. 2/
What is true is that the market is expressing a high degree of confidence that inflation has peaked, and the Oct CPI slowing is the start of a trend. That view could be wrong of course, but the market isn't crazy for drawing such a conclusion. 3/
One fascinating question in the next few years is whether the so-called "term premium" will return to the U.S. Treasury market. This is amt of extra yield you get paid simply for holding longer maturities independent of your future expectations. Some thoughts:
There isn't much people agree on IRT the term premium, but there is general agreement that is seemed to decline in the post-GFC period. IMV, the reason *why* it declined is the key to determining if it might come back going forward. 2/
On thing that complicates analysis of the term premium is that it is very hard to measure. The true term premium is independent of expectations. I.e., if the curve is steep bc people expect the Fed to hike in the future, that's not term premium, that's expectations. 3/
Lost in all of the "pain" talk from Powell yesterday, there was some interesting signals about their reaction function. A🧵on my takes:
First let's remember that Powell's *current* job is to talk as tough as he can. They can't risk markets thinking that falling stocks or signs of a recession will cause them to lose their nerve. I don't think we get much signal from his tone right now. 1/
Powell also doesn't really know how much of a slowdown it will take for inflation to subside, or how quickly it will fall once it starts falling. If he lets up now and inflation proves more stubborn, it risks the Fed's price stability rep. He can't take that risk. 2/
In macro, we never get definitive information from any data point. There's random noise, measurement problems, seasonality, etc.
OTOH, investors can't wait until there is enough data for absolute proof. The mkt will have long ago moved. Ergo we have to operate on the margins. 1/
Here's one way to think about it. Say I handed you a die and just told you to roll it. You roll a 6. Would you draw any inference there? Of course not. A 6 is just as likely as any other number. 2/
But what if I told you there's a *chance* the die is loaded to roll 6 more often. I don't say what the chance is nor how loaded the die might be. All you know is there's a *possibility* that a 6 isn't a coincidence. 3/