• Month-end means bond traders are adjusting portfolios to match a published index
• The index was a bit less friendly than expected yesterday, and bonds bounced
• That bounce gives us a good target to watch for rallies

Things have been going fairly well for bond markets over the past few days, with Tuesday's post-Yellen rally being the biggest potential turning point.  Before that, Treasuries and MBS had been in a mostly-sideways pattern, potentially running into resistance at middle-of-the-road trading levels.  But after the Yellen-inspired rally, we've had a hard time making new progress.  This begs the question: did month-end bond buyers simply use the Yellen speech as justification to buy early and will we now have a tougher time making additional gains?

We've discussed the effects of month-end bond buying quite a lot over the years.  In most discussions on the topic (here or elsewhere) there's a risk or tendency to oversimplify the phenomenon by saying that "more bond buying occurs at the end of the month."  Such an oversimplification isn't necessarily a bad thing, because it's largely true.  Traders really do have some extra bond-buying needs at the end of the month, but it's not quite as simple as it sounds.

Much of the month-end bond buying is more of a re-balancing of duration.  In other words, if traders add up all their bond holdings and weight them according to maturity, there's an average duration.  In other words, if I only owned an equal amount of 5 and 10 year Treasuries (and nothing else), my bond holdings would have a 7.5 year duration.

Every month, Barclays publishes an index that tells fund managers what duration they need to hit in order to be keeping pace with the overall bond market (incidentally, they break it down into sectors like Treasury, Agency, MBS, etc.).  The reason we're almost always talking about month-end bond BUYING is that the index is almost always extending.  The reason for this is simple: all things being equal, bonds are always maturing and getting paid-off, so you will always need to buy more if you want to maintain a constant level of duration over time.

For instance, in my example above about a 7.5yr duration, in 6 months from now, if I didn't buy any more bonds, I would only have a 7yr average duration.  My bond holdings are being repaid!  But if the broader bond market continues trading perfectly sideways, I'd have fallen 6 months behind the average duration, and would need to buy to get back in line.

The point is that the bar is set pretty high for bond market weakness to push the Barclays index into negative territory.  It hardly ever happens for Treasuries and traders know that.  They also have a pretty good idea of where the index will come out.  It's when the actual index differs from the expectation that it can have a noticeable effect on trading.  This month's final numbers were out yesterday and came in with a bit smaller of an extension than traders were expecting (Barclay's estimates came out on March 23rd, and many traders figured that the average duration would extend a tiny bit more on top of the initial estimate, but yesterday morning's final numbers showed no change).

This could help explain some of the weakness in the longer end of the yield curve that we saw in the immediate wake of the update.  It's complicated, of course, by the presence of Yellen's speech the day before, which also had the effect of benefiting shorter duration bonds.  At the very least though, the levels intact just before yesterday morning's selling provide a good rally target (these are the overnight lows in 10yr yields at 1.80).  

As always, expect a greater chance of volatility around 3pm today.  In addition to being month-end, it's also quarter-end, which means even more traders are adjusting their holdings for book-keeping purposes (i.e. the movement might not necessarily have anything to do with news or events.  Traders could just have trades to make, whether or not it makes sense from a fundamental or technical standpoint).