***REQUIRED READING for anyone who doesn't fully understand why the 4.5 UMBS coupon is far better to follow than the 5.0 intraday (for now).***

The morning commentary is typically focused on market movement and broader bond market considerations for the rest of the day. Today, however, there's an urgent need to discuss an often confusing topic when it comes to the bond market's relationship with mortgage rates: which MBS coupon should you be watching? 

In the past, for the most part, that's been easy for a loan originator to determine based on the rates being quoted (as long as the originator understands that MBS coupons can accommodate rates that are up to 1.125% above the coupon).  Prevailing rates consequently suggest MBS coupons of 5.0 and 5.5, but these are absolutely NOT the coupons that should be followed for the purpose of identifying intraday trends and risks.  

You may be wondering, "if a 5.5 coupon is the best match for what I'm seeing on my rate sheets (i.e. rates of 5.75-6.625%), wouldn't it be a no-brainer to watch 5.5s?"

Your logic is sound, but there's a big problem right now: compared to lower coupons, 5.5s are effectively nonexistent.  There's no depth from previous trades and new day-to-day volume / liquidity is best described as pitiful.  The same is true for 5.0 coupons.  In fact, the same was even true for 4.5 coupons until very recently.

"B-b-b-but..." you may ask, "if 5.5 coupons are just right, and 5.0 coupons are arguably just a bit too low, sure it has to be better to watch 5.0 coupons than going even lower to the 4.5?!"

NO IT'S NOT!  In fact, 4.0 coupons were better than 4.5 coupons up until the past few hours, and 3.5 coupons were better than 4.0 coupons throughout April, even though we were quoting rates over 5% (3.5 coupon maxes out at rates of 4.625%).

What do I mean by "better?"  Let's put it this way.  At any given moment when you're looking at an intraday MBS price, there's a certain quality/accuracy/reality associated with that price based on how close it comes to the prices where trades are actually occurring.  Intraday pricing feeds generate A LOT of quote activity that doesn't necessarily correspond with actual volume.  When buy/sell (aka "bid/ask") quotes are close to one another, that's what we would consider a high quality price and a liquid market.

Now chew on this fresh example:   From 8:11am to 8:50am ET this morning, the bid/ask spread on 4.0 and 4.5 coupons moved in a range from 0.06 to .375. Over the same time frame, 5.0 coupon bid/ask never dropped below .625. There's no great way to be emphatic enough about how completely worthless MBS pricing is when the bid/ask is .625. Even the .375 in the other coupons is laughably worthless.  Point being, if you were watching 5.0 coupons this morning (or yesterday at around 3-330pm), you were watching an inferior representation of how the mortgage bond market was actually moving (EVEN THOUGH 5.0 coupons would be a better baseline/benchmark when setting rate sheets).

The following chart shows the bid/ask spread referenced above.  Note the sustained blowouts in 5.0 coupons:

That chart is a bit messy because it follows each tick of data.  The following chart uses moving averages of the same data to smooth things out quite a bit (so much so that we can even bring 3.5 coupons into the mix to see how liquidity has evolved over the past few months).  Definitely be sure to note that yesterday was the first time that 4.5 coupons were as liquid as 4.0 coupons.  Also be sure to note that yesterday was the day that I changed our target coupon to 4.5 from 4.0 (THERE IS A METHOD TO WHAT SOME OF YOU PERCEIVE AS MADNESS!).

As you can see, the only true madness at this point would be to follow 5.0 coupons as your guiding light for intraday risks/trends.  To help bring this all home and add some institutional-level perspective, here's what your friend and mine, Adam Quinones has to say about all of the above (with my explanatory notes in parentheses in the few places I think it might be needed):

Predicting reprices in this environment is far more challenging because all liquid coupons are priced below par. Rate sheet generation complexity only increases when you look at coupons priced above par. These MBS are not liquid because there is not enough outstanding supply of those mortgages. It's a chicken or the egg dilemma. The lack of liquidity in premium (over PAR) MBS coupons forces secondary marketing departments to take extra steps to ensure they're building rate sheets with a benchmark (the "guiding light" from the previous paragraph) that reflects the actual price they can sell the bond for when it's ready for shipping. Because the screen doesn't always reflect a real market for those premium coupons, secondary has to call broker-dealers to get quotes. These quotes are not stable. Dealers are basically pricing them on spread to other coupons and adding a risk premium. The end result is high uncertainty and unpredictable reprice risk.

If it's not 100% clear, what AQ is saying is that the actual foundation/benchmark/touchstone/team captain/etc for day to day MBS movement is the coupon that is the most liquid and/or the most heavily traded, but that those coupons don't currently align with rate sheets.  As such, the secondary market uses the prices from those liquid coupons and runs some math in order to generate a good guess at where higher coupons should be priced.  It's half art, half science, and definitely prone to imperfections on both fronts.

The super duper bottom line is this: 

If you were ONLY trying to create a rate sheet for the day, yes, you would need to have an idea of where 5.0 and 5.5 coupons were priced.  But if you tune in to intraday MBS prices for the purpose of following trends in rates and reprice risk, it's the lower, more liquid coupons that matter.  You can ALWAYS trust that I am watching this like a hawk so you don't have to.  I will always put the gold star next to the coupon of choice.