From an analytical standpoint, today's trading session was not well-behaved or easy to assess (I saw several iterations of "no one has any idea what happened today" around analysts' campfires).  Bonds began the day in relatively unchanged territory after experiencing essentially no overnight volatility or drama.  The opening hours were calm.  What's especially striking about that is the absence of volatility surrounding the 8:20am CME open.  Reason being: I'd typically want to see some 8:20am volatility in order to make the conclusion that "new month" tradeflows are driving a rally or sell-off.

To understand "new month" tradeflows, first consider the month-end bond market environment.  By the end of the month a certain portion of bond traders MUST be holding a certain a mix of bonds in order for their portfolios to adhere to certain indices.  In other words, if a specific retirement fund advertises "40% US government bonds," there will be some fine print saying "US government bonds" will be allocated according to an index such as Barclays Aggregate Treasury Index.  That index provides an average duration and that has to be matched by the average duration of the money manager's portfolio. 

Bottom line: month-end creates some constraints for bond traders.  They are compelled to hold certain positions and can then trade more freely at the start of the following month.  This will occasionally result in the first day of any given month seeing a spike in volume and momentum that is apparently unrelated to other motivations.  Today was definitely that sort of day.  There were no other great ways to account for the 7+bp rally in 10yr yields or the 3/8th point gain in Fannie 3.5 MBS.  

I think a lot of analysts will try to draw some connection between the 10am ISM data and the rally, simply because the timing lines up fairly well.  The problem is that the volume definitely does NOT line up.  Moreover, it's really a stretch to say that 57.2 vs 57.0 in ISM (not to mention the higher inflation component) amounts to a compelling case for a bond rally.  If anything, the traders who were inclined to buy bonds from the get-go were simply waiting to make sure the ISM data wouldn't offer a major objection to the idea.