GS US SEMI EQUIP B-T-B: BACK-END
WEAKNESS AND SIGNIFICANT CAPACITY
ADDITIONS
Summary: SEMI released the July US book-to-bill (b-t-b) ratio of 1.05 (GS and the Street
1.07). Orders were flat m-m (1% below our estimate and in- line with the Street) and
shipments were +2% m-m (1% above our estimate and 2% above the Street) - higher than
expected shipments (denominator) drove the overall ratio lower. Back-end orders declined
10% m-o-m and were light of our and the Street estimate driving the back- end
book-to-bill ratio below 1.0 for the first time in 18 months. We continue to believe that the
back-end typically leads the front-end into and out of downturns, and recent datapoints
suggesting front-end order weakness support this thesis. Data also shows significant
capacity additions as yoy increase in front-end shipments was 2nd highest in history. No
change to our belief that investors should continue to avoid overweight stock positions as
we believe that the industry is in the early stages of a downturn which is not yet priced
into the stocks.
3-MONTH ROLLING AVERAGE BOOK-TO-BILL OF 1.05 WAS 0.02 BELOW OUR
1.07 ESTIMATE ON IN-LINE ORDERS AND HIGHER SHIPMENTS (HIGHER
DENOMINATOR DRIVES THE OVERALL RATIO LOWER). Semiconductor
Equipment and Materials International (SEMI), a semiconductor equipment industry
trade association, reported its three-month rolling average July bookings and shipments
statistics on Thursday night. The U.S.-based semi equipment suppliers' book- to-bill ratio
was 1.05, 0.02 below the GS and Street estimate of 1.07. Orders were $1,610 million (flat
month-over-month and +128% year-over- year), 1% below our estimate of $1,630 million
and in-line with the Street estimate of $1,610. Shipments were $1,538 million (+2%
month-over-month and +96% year-over-year), 1% above our estimate of $1,530 million
and 2% above the Street estimate. Recall that a higher than expected denominator
(shipments) drives the overall ratio lower. The front-end book-to-bill was 1.07 on orders
+3% month-over-month and +157% year-over-year and shipments + 2%
month-over-month and +102% year-over-year. The back-end book-to-bill ratio was 0.95
on orders -10% month-over-month and +51% year-over-year and shipments +3%
month-over-month and +74% year-over-year. The back-end book- to-bill ratio of 0.95
was the first time in 18 months that the ratio was below 1.0. June's overall book-to-bill
ratio was revised downward to 1.07 from 1.08 on 2% higher shipments (again, recall that
the higher denominator drives the overall ratio lower).
GS: BACK-END ORDER DECLINE OF 10% M-O-M SUPPORTS OUR BELIEF THAT
THE INDUSTRY IS IN THE EARLY STAGES OF THE DOWNTURN. Back-end
orders declined 10% month- over-month to $299 million in July and were light of our and
the Street's estimates (GS $330 million and the Street $316 million). July marks the third
consecutive month of a decline in back-end orders and also marks the most pronounced
decline in back-end orders since August of 2002. The weak back-end order number is
in-line with recent reports from Agilent (which reported a 27% sequential decline in
semi-test bookings in its most recent quarter) and Advantest (which guided for a 20%
sequential decline in bookings in CQ3) as well as with recent commentary from other
back-end equipment vendors.
We continue to be believers of the notion that the back-end typically leads the front-end into and
out of upturns/downturns, which we believe is supported by the book-to-bill data. For example,
back-end orders led the front-end into the downturn in 1995, as back-end order growth began to
slow in April of 1995 and then began to decline month-over-month in May of 1995 while
front-end orders remained robust until July of 1995. During the 2000 cycle, the back-end began
registering month-over-month declines in orders in April of 2000 while front-end orders continued
to increase month-over- month until November of 2000 before falling off significantly. During the
head-fake upturn that began in late 2001, back-end orders began to improve significantly
month-over-month in December of 2001 while front-end orders began to improve two months later
in February of 2002. Back-end orders also led the front-end out of the head-fake upturn in 2002,
declining month-over- month in June of 2002 while front-end orders only began to decline monthover-
month in August of 2002. Back-end orders also led the front-end into the upturn in
2003-2004, as they began to improve month-over-month in January 2003 while front-end orders
only began to improve month-over-month in August of 2003. Given the history of the back-end
leading the front-end during turns in the cycle, we would expect front-end orders to show
incremental weakness over the coming months. We believe that this is supported by increased talk
among the front-end equipment companies (KLA, Varian, Asyst, and Brooks) of customer
push-outs and cancellations.
2ND HIGHEST YEAR-OVER-YEAR INCREASE IN FRONT-END SHIPMENTS SUPPORTS
THE BELIEF THAT THE SEMI INDUSTRY IS SET TO ADD SIGNIFICANT CAPACITY IN
2005, CALLING INTO QUESTION THE IDEA OF A LESS SEVERE DOWNTURN. We
believe that Thursday night's data calls into question the notion of a less severe downturn. Many
investors are making what we believe to be a reasonable argument that the semi industry has not
added as much excess capacity in this cycle (due to a less robust upturn) and that the next (current)
downturn will not be as severe as previous downturns. We believe that the front-end shipment data
provided by SEMI argues otherwise. Front-end shipments increased 102% year-over-year, the
second highest year-over-year increase in the data set. While many clients will argue that this
increase is off of a low base from one-year ago, we would highlight that the starting base from one
year ago is almost exactly equal to the trough shipment level in the previous downturn. We would
therefore argue that the 102% y-o-y increase in shipments supports our recent supply/demand
model which argues that the semi industry has already added significant supply that will drive
utilization rates down significantly in 2005 and lead to lower semi equipment order rates over the
coming quarters.
AUGUST BOOK-TO-BILL EXPECTED TO DECLINE TO 1.03 ON A 2% M-O-M DECLINE IN
ORDERS AND FLAT SHIPMENTS. We are estimating a decline in the book-to-bill ratio in
August to 1.03 on a 2% m-o-m decline in orders and flat shipments. We are modeling three-month
rolling average overall orders of $1,575 million (-2% month-over-month) and overall shipments of
$1,535 million (flat month-over-month). We estimate front-end shipments of $1,225 million (flat
month-over-month) and front-end orders of $1,290 million (-2% month- over-month), yielding an
estimated front-end book-to-bill ratio of 1.05. We estimate back-end shipments of $310 million
(-1% month-over-month) and back- end orders of $285 million (-5% month-over-month), yielding
a back-end book- to-bill ratio of 0.92.
WE CONTINUE TO RECOMMEND THAT INVESTORS AVOID OVERWEIGHT STOCK
POSITIONS AS WE BELIEVE THAT THE INDUSTRY IS IN THE EARLY STAGES OF THE
DOWNTURN. We continue to highlight that the book-to-bill should not be a significant trading
event for the stocks given that it is a backward looking and unaudited metric. We continue to
recommend that investors avoid overweight stock positions as we believe that the industry is in the
early stages of the downturn (as supported by the month-over-month decline in back-end orders
reported in the July book-to-bill data).
I, Jim Covello, hereby certify that all of the views expressed in this repor |
