The bond markets took the latest Fed OMC announcements as an "all-clear" sign to buy Treasuries.  The move has pushed the Treasury yield curve a lot flatter.  As of mid-day Monday, the 2-10 spread was inside of +160, closing in on its one-year low of ~+150.  The 5-10 spread has also narrowed, although at +109 it's still about 20 basis points above its 12-month wide.  The reaction is pretty simple: if the Fed (the biggest factor in the markets) has said that price risk isn't a problem, why not pick up the extra carry by owning the 10-year?  (A sobering thought is that the 10-30 spread has widened out since the Fed announcement-there still is some caution in the market after all.)

One thing that has helped the MBS markets since the beginning of the year is the decline in volatility.  Implied swaption vols are at their lowest levels in over a year.  For those unfamiliar with the terms, "implied volatility" is just another way to quote the price of an option-think of it as a more universal way to quote option values, rather than using the options price (or, technically, its "premium").  In other words, option prices have been trending lower since last fall, and have recently moved sharply lower.  This may be the result of the downtick in actual (or "realized") volatility.  The 10-year Treasury, for example, has traded in roughly a 35 basis point channel since the fall.

Most of this was noted in a recent MBS Commentary post, but now, let's explore why a downtick in volatility helps MBS and mortgage assets.  Simply put, any mortgage-related asset can be viewed as a position combining 1) a long position in a cash flow, and 2) a short position in an option.  If the option becomes more valuable, the overall bond is worth less; conversely, if the option is worth less, the value of the MBS bond increases, all else equal.  It's not a lock-step reaction; however, this works its way into both portfolio management and the valuation tools used by investors, and is one reason that MBS have performed so well in 2012.

Last time, I introduced the concept of "dollar rolls" and ways to think of them.  The concept of rolls is very important to secondary market managers, since it strongly influences what settlement month should be sold.  The best way to think of dollar rolls is as a choice of what you'd want to do at two points in time, based on the future value of the different cash flows.  The trader can look at the roll value two ways-either how much better is the roll trading than implied by the net cash flows (i.e., what is the market drop versus the amount calculated using the investors inputs) or what is the break-even cost of funding using the actual market drop as compared to my actual cost of funds.  A better value for the roll (i.e., either a bigger implied drop or a lower break-even cost of funds) means that the roll is trading "special." 

Let's take a brief look at some numbers.  Good-day settlement in February for Fannie 3.5s is 2/13, while settlement in March is 3/12.  Assume that the market is trading as follows:

FN 3.5s in Feb:  103-18;

FN 3.5s in Mar:  103-9 (i.e., a 9-tick drop).

Further assume a 3% CPR prepayment speed and a cost of funds of 0.50%.   If I roll the cash flows (i.e., buy them for the back month) I have the following cash inflows:

If I buy them for the front month my cash flows are as follows:

In this case, I have an advantage of $537 if I roll the bonds versus if I hold them (and accept the P&I cash flows.)  In price terms, this equates to 1.72 ticks, which means the roll is trading a little more than 1+ special.  If I want to express it as cost-of-funds, my reinvestment income (i.e., what I earn on the money I haven't had to use to pay for the securities) comes out to -$133, if I were to break even to the $1,037,908 proceeds for buying and holding the securities.  This means that my cost of funds is -0.167%.

There are a number of factors that impact the roll market.  A lot of buying in the front month (to settle CMO deals, for example) will tend to widen the roll and make it trade special, as will a lot of selling of the back month by originators.  Rolls will often get special around the front-month settlement, as trading desks that are short the front month (and need to make deliveries) will need to buy the roll.

As a secondary market manager who is typically shorting into the market, selling the later months is essentially "selling the roll."  If a roll is trading special, you'd want to think about selling the earlier month (i.e, selling at the higher price) and rolling the hedge forward around settlement date.