It might seem cliche, but everything really does come down to NFP tomorrow, and by "everything," I mean the next 6 weeks... In case you missed last night's closing commentary, there was a chart that showed a very strong similarity in MBS movements over the past 3 years compared to this year. In fact, on the exact same weeks in history, annual lows were made in MBS four years in a row. On the exact same week in history, prices began to weaken and head toward the only only other major low before eventually moving higher into the winter. And you guessed it, those low points also occurred on the exact same week! Put to picture, it goes something like this:
Why are we talking about this two days in a row? Simply put the yellow circles in the above chart are all the first week of September. In case you don't have a calendar, we are currently in the first week in September. So it may be worth considering any other factors that would bolster or refute the above synchronicity. For the record, the synchronicity makes a case for the next 6 weeks or so being sideways or negative, beginning now.
On the refutation side we have the simple and obvious fact that if these similarities had occurred 4 or more years in a row as opposed to the three shown above, we would have opted to show you that. So as you may have already inferred, looking at the next year back refutes the correlation. That's not to say, however, that the pattern doesn't exist at other points in the past. In fact, in over 10 of the past 15 years prices have bottomed in summer and peaked in the winter. Perhaps the best broad thesis would be to say that price peaks tend to fall in the colder half of the year and lows tend to fall in the warmer half. Some of the exceptions are glaring, but in a general sense, this is the way it is.
Another point of refutation would be the exceptional nature of 2009. In other words, when making a case for a reasonably high degree of repitition in price movements, one would probably assume that the minority of data that does not adhere to those similarities is most likely created during years that are UNLIKE the average year. I suppose there are plenty of different metrics for determining that LIKENESS, but I hope you won't need them enumerated in order to regard 2009 as unique.
On a sort of tangent, part of 2009's uniqueness lies in some of the "all time" levels for data, such as tsy futures hitting 130-00. Like I said yesterday, this isn't likely to happen until Great Depression part 3. So perhaps it should calm our urges to predict the future with the past simply knowing that we're at a significantly different starting point this fall. But at the risk of making you dizzy trying to figure out if we're talking about refutation or support, a counterpoint to that counterpoint might be the fact that 2006 was a very flat year compared to 2008, yet the general descriptions of the "circles" in the chart above can apply to both years. That suggests that the magnitude of price changes doesn't matter as much as the date ranges...
In discussing the support for the correlation--not a scenario that is going to benefit rates in the short term--consider what you've just read in conjunction with the following day-over-day chart updated to include today's close:
The support is a much more simple discussion... Though we didn't have the pretty horizontal lines on the chart last night, we essentially posted that above chart in yesterday's close. This was to discuss the approach of both MBS and tsy prices to their absolute best levels of the entire summer. Bouncing or crossing at such levels is informative regarding impending trends, or more appropriately, the conclusion or continuance of previous trends. With yesterday's momentum, the default perception seemed to favor "strong bonds, weak stocks." I tended to agree until my afternoon research yielded the "circle chart" in conjunction with the re-test of summertime's best levels. It seemed to me then, as it does now, that the result of this retest of the above horizontal lines would either confirm or reject the broader hypothesis from the top of this commentary.
Well... Yeah... That's pretty much it... We DID NOT break through either line today, and perhaps even more importantly have FAILED to hold above a very similar line in tsy futures. In other words, we're back in the fiery pit (fyi, the chart below has not been updated with today's prices, but we've fallen back into the pit.):
None of the above mentioned failures to break or confirm are inherently positive or negative for MBS. They are merely "NOT POSITIVE."
And so it is that NFP gets it's oft-assigned cliche status as a market mover. If prices do continue to come down tomorrow, will we ever know if it was due to NFP or some mystical technical forces working behind the scenes? I don't think KNOWING matters so much as preparing for different outcomes. I'm never fond of predicting the future, but in a vacuum of data and news, I think the default movement in bonds tomorrow is negative. In other words, if NFP falls within a certain deviation from consensus, regardless of whether it's better or worse, bonds lose. The exception would be the unknown quantity that is stock market sentiment. Perhaps equities will need more convincing that an "as expected" NFP print in order to restart their rally. If stocks get involved in some runaway declines, that could be enough to prop up the bond market for another day.
And that would get us through until next week when there is a whole new round of "turning point" style data including digestion of MBS prepays, Class A settlement, and the end of summer vacations among other things.
MBS Market Alerts
