SlowThinker, I gave some thought to the factual content of the blurb.
I'll assume Niles is the author of the note since he is their analyst
covering the stock.
"Believe Q2 could be much better than anticipated. Believe that total
revenues could be up almost 50% sequentially due to bits up 40-50%
and up sequential pricing".
Due to bits up 40-50%. Since that "due to" ties back to "revenues",
he's talking shipments. Obviously, if it took a New York Times
reporter to alert him to some of the interesting things in the recent
10-Q, Q2 certainly I can see why he would "Believe Q2 could be much
better than anticipated."
I could leave it right there, but then I might be accused of the
dreaded "ad hominem" attack <g> (I'd plead "not guilty" as IMHO, his
position as a professional research analyst at a major i-bank
inextricably links him to his work product), but here's where I'm
coming from...
I have the 12-24-98 Niles report.
His estimate for Total DRAM Bytes (mil.) for 2Q was 492.
The "Total DRAM Bytes" row title is coming under a column heading
that reads "DRAM shipments".
He has them down for 1Q coming in at 394.
Based on those numbers he has them coming in at 25% sequential growth
for 2Q.
Sequential growth 4Q to 1Q is down as was +7%.
The 10-Q states that total bit shipments declined 10%.
So let's go back and look at what he had down for 4Q, make the
adjustment and see where that puts us.
For 4Q he has them down as 368.
So, for 1Q, if shipments were down 10%, that would be 331.2.
So, he was projecting 492 for 2Q'99 on a 12-24-98 doc.
He was viewing it then as a 25% sequential increase.
Using his 4Q'98 numbers and adjusting them to reflect the 10%
decrease in shipments cited on the 10-Q, the 1Q figure now comes in
at 331.2.
Take a wild guess what the sequential increase is?
48.55%
I must note that the NYT piece of last weekend certainly made it
seem as if his familiarity with the 10-Q was, shall we say,
"limited".
And he's saying what now?
Believe Q2 could be much better than anticipated. Believe
that total revenues could be up almost 50% sequentially due to bits
up 40-50% and up sequential pricing.
I guess if you thought the sequential growth was 25% based on a
faulty baseline, and when you adjust the baseline you get 48.55%, you
would be under the belief that "Q2 could be much better than
anticipated".
Based on the available information I'm looking at:
A. His 12-24-98 report.
B. The MU 10-Q.
C. The NYT article.
That's my line of reasoning.
The possible flaws are
A. he may not have gone back and corrected his own work to reflect
the 10% decrease in shipments, and
B. he may actually have a different in their for 2Q shipments.
I signed up for E*Trade's professional edge trial and when I checked
earlier, it wasn't there. I'll certainly be interested in seeing how
he dealt with his errors and what the numbers are. While
acknowledging the possibility of flaws, I'd have to say that right
now I suspect he probably dealt with it exactly as I thought he would.
However, if you choose to view it as an ad hominem attack (and I
believe your strongest ground is arguing it is the "circumstantial"
ad hominem) and wish to successfully rebut it, you will have to
prove that the truth or falsity of the proposition (his belief that
2Q will be much better than anticipated, in part, "due to bits up
40-50%") has nothing to do with the circumstances of the person
(presumably, Dan Niles) making the proposition.
IMHO, that's difficult based on the above line of reasoning; however,
if he changed his total bit number for Q2 maybe you've got a little
daylight there <g>
As far as "total revenues", while I can see how that's possible, I'm
with you on questioning the likelihood.
Assume the total memory represented by 64Mb shipments will increase
40% sequentially.
Assume their ASP's are 9.75.
What has to happened for total revenues for semicon memory products
to grow 50% sequentially?
64 1Q % change 2Q(E)
units 29.707 41.590
asp 8.96 9.75
revs 266.175 52.34% 405.501
memory 1901250 40% 2661750
non-64
revs. 143.325 46% 208.7490234
TOTAL 409.5 50.00% 614.25
(Source for 8.96 1Q 64Mb ASP, 12-24-98 Niles report; source for 9.75,
what I'm sitting on).
My problem with the above is the following:
If the 64Mb is their bread and butter product and it's also a product
that is a larger % of total sales, what's happening out there that's
going to drive non-64 revs. to a 46% sequential increase?
Granted, the higher you choose to assume the 64Mb bit growth will be,
the lower those non-64 revs have to be.
Good trading,
Tom |
