Tuesday, October 5, 2010

Friday's September Employment Report

To say that that there is a lot riding on the September employment report is
probably an understatement, as it clearly has the potential to settle the issue of
whether the Fed will proceed with another round of quantitative easing in a few

A number of things to be kept in mind for Friday.

1) In the establishment survey, the emphasis, once again, should remain on private
payrolls, as the unwinding of the Census workers will continue to distort the headline
payroll print. The "Census effect" may not disappear until December, although it should
be gradually fading in the October-November period. 

2) Private payrolls have averaged 95K a month since the beginning of the year, reflecting
the creation of a total of 763K jobs in the private sector since January. More than half of
these jobs were created in just two months in early spring (March-April), prior to the onset
of the "soft patch". In the last 4 months, the average monthly gain has slipped to 71K,
representing a total of 286K jobs created in that period. 

3) The question of what number is needed to make the Fed willing to at least temporarily
suspend the activation of new QE measures requires a high degree of oversimplification,
as the Fed will not be making that decision based on a single number in the employment
report. With that in mind, a gain of over 130-140K would probably help the debate alive
within the FOMC as to the necessity for such action. Of course, revisions to the prior two
months will also matter, as any surprising strength in the September private payroll
number could be offset by revisions in the opposite direction.

4) Speaking of revisions: Along with the September employment report the BLS will also
be releasing its benchmark revision to the level of nonfarm payrolls in March 2010. This is
the BLS's first estimate, with the final number to be released along with the January 2011
employment report next February. There is no way of predicting whether the level of
payrolls for this past March will be revised up or down, but there seems to be growing
chatter in recent days that the number will be revised downward- by a potentially
appreciable amount. (What counts as "appreciable" with regard to the benchmark revision
is probably something to the tune of 300-400K or more). This is based on the impression
that the BLS has a bias toward overestimating job creation in periods of weak economic
activity. Also, another factor contributing to the talk of a substantial downward revision to
the March payroll level is the significant discrepancy between the ADP survey and the
BLS data over the last year - with the former showing a much slower pace of job growth.
With that taken into account, it needs to be emphasized that we have no hard information
pointing safely toward a sizable downward revision to the benchmark March number.
(For the record, the biggest downward revision in recent memory was the one announced
last fall for March 2009, which was by more than 820K).

5) There seems to be an unusually strong consensus that the unemployment rate will be
up in September to 9.7% from 9.6% in August. There are two main reasons behind that
expectation: 1) the civilian labor force showed a sharp contraction in the June-July period,
shrinking by a cumulative 833K. It is fairly uncharacteristic to see a contraction of the
civilian labor force in a phase of the cycle where, on balance, the economic
environment is improving compared to a year ago. Although the civilian labor force
bounced back by 550K in August, there is still a deficit of 285K jobs over the last three
months, which most analysts expect to be rectified in September in the form of a jump in
the size of the civilian labor force. 2) The unemployment rate in August was actually
9.64%, which, in conjunction with the above point about the labor force, makes it a natural
to inch to the 9.7% mark. 

While the unemployment rate is largely a lagging indicator and unlikely to change
materially anyone's perception of the broader economic environment, from the Treasury
market's standpoint a potential spike above the widely anticipated 9.7% (and with a little
help from an "s advertised" private payroll number) would be viewed as a license to
continue its seemingly irrepressible rally.

Anthony Karydakis

About the Fed's Purchases of Treasuries

The Head of the NY Fed's Open-Market Trading Desk, Brian Sack, gave a
thoughtful speech yesterday at the CFA Annual Conference, where he
discussed the Fed's decision on August 10th to reinvest the principal
payments of maturing agency and MBS debt. He also discussed at some
length the various issues associated with the prospect of any further
expansion of the Fed's balance sheet. The full text of his talk can be
found at


It is worth keeping in mind that the Fed is already committed (before
they consider any additional such measures in the coming weeks) to
purchasing a very sizable amount of Treasuries over the medium-term, as
a result of the August 10th decision. At the time, they estimated that
the amount of agency and MBS debt that would be running off by the end
of 2011 was likely to be around $400 billion. In his speech yesterday,
Brian Sack stated that this estimate now is already"somewhat higher"
compared to what was considered on August 10th. This is the direct
result of the strong rally in the Treasury market in the interim, with
10-year yields declining by approximately 35 basis points since August
9th. (The Treasury rally and the ongoing tightness of mortgage spreads
increases the pace of MBS prepayments in the Fed's portfolio).

If long-term Treasury yields were to decline moderately further in the
near future, the Fed's "automatic" purchases of Treasuries would
increase even more. Of course, the exact additional amount of Treasuries
that would need to be purchased over the next 15 months (over that
initial $400 billion the FOMC mentioned on August 10) would also be
partly contingent upon the length of period that rates remain close to
their lows over that time frame. However, with the economic recovery now
expected to remain on a very moderate growth path, yields are likely to
remain low well into 2011.

All in all, the Fed may already be on track to purchase $500 to $600 billion of
Treasuries by the end of next year, WITHOUT explicitly announcing any additional
QE next month. Now, if the Fed does announce a stand-alone, "active" program of
additional Treasury purchases in November- and that adds significant
further momentum to the Treasury rally, then the combined effect of the
two programs of such purchases will be magnified due to an even faster
prepayment rate of mortgages as yields continue to decline. The end
result here is that the Fed may end up buying an appreciably higher
amount of Treasuries over the next 12 to 15 months than officially
stated; depending on the trajectory of economic activity over that
horizon, such a number could easily exceed the $1.5 trillion mark.

Anthony Karydakis