We've definitely seen Treasuries improve relative to MBS more often than not in the past week or two.  The most obvious driving force behind that phenomenon was fairly simple.  It had to do with MBS outperforming even more noticeably in the 6 weeks before that (a feat which was incidentally attributable to MBS severely underperforming during the initial coronavirus volatility.  In other words, MBS did so much better vs Treasuries that it made sense for the relationship to equalize a bit.

Another major consideration is and has been the Treasury supply outlook.  Fiscal stimulus and anticipated revenue shortfalls create the need for big increases in Treasury auction amounts.  That creates Treasury-specific weakness.  10yr yields aren't necessarily thinking about this as traders gear up for the start of this week's Treasury auction cycle today (2yr notes at 1pm), but it contributes to Treasury-specific weakness in general.  And yes, this can occasionally be seen in some random selling pressure in the days and hours leading up to the start of a new auction cycle EVEN THOUGH those auction amounts and times have been known for weeks.

More important than dissecting the MBS vs Treasury performance today is the simple task of remembering that the broader trend is not our friend, even though it hasn't been much of an enemy so far.  In fact, one might argue that any gently-sloped trend that suggests 10yr yields remaining under 1% for several more months is about as friendly as we could hope for.  While I wouldn't expect this narrow range marked by the yellow lines below to continue unabated until hitting 1% this fall, it's the best way to illustrate the market's gradual recovery from coronavirus impacts for now.