News & Analysis
Comment
Philosopher2
PremierSnow
It's nothing new, they retroactively change monthly estimates all the time.
WSTS strips value from Q1 chip market
Peter Clarke
6/3/2011 12:38 PM EDT
LONDON – The World Semiconductor Trade Statistics organization has reduced its figures for global chips sales for each of the months of January, February and March of 2011, in its latest updating of its spreadsheet.
Latest actual global semiconductor sales figure for January 2011 is $25.173724 billion compared with $25.183883 billion published previously; for February the new figure is $22.164019 billion versus $22.197153 billion; March sales are set at $28.337218 billion against an earlier figure of $28.384976 billion.
The net result has been to strip about 91.1 million in value out of the first quarter's sales figures, which is now listed by WSTS at $75.675 billion, compared with the previous figure of $75.766 billion.
The changes also affect the three-month average figures produced by WSTS and which are used by the Semiconductor Industry Association and the European Semiconductor Industry Association. The SIA and the ESIA use the averaged sales figures to smooth out the monthly data that would otherwise show troughs at the beginnings of the quarters and peaks at the ends of the quarters.
The SIA appears to have taken the latest reduced figures into account in its April announcement.
WSTS is a non-profit organization of 64 semiconductor companies representing more than 70 percent of the world semiconductor market, and is widely regarded as the most accurate source of data on the global and regional semiconductor markets.
Related links and articles:
www.wsts.org
News articles:
Japan 'actual' chip sales collapsed in April
April global chip sales show softness
WSTS cuts $1 bn in chip sales, says analyst
|
|
|
|
|
Navigate to related information
Warren
6/3/2011 4:45 PM EDT
Eight significant digits. That's, um, quite something. 91 million dollars is nothing to sneeze at I know, but this update points out a 0.1% adjustment for first quarter sales. I wonder if many folks visit EETimes looking for adjusted sales figures where the "news" is that previous estimates were within 0.1% of current estimates.
Sign in to Reply
peter.clarke
6/4/2011 6:10 AM EDT
@Warren,
Your point is well made.
The signifcant digits are those of WSTS.
But given that so much analysis is based on the numbers given by WSTS I thought it was worth reporting.
There is also a tendency for numbers to get decreased in retrospect, but never increased.
This would have the effect that the monthly data published by the SIA and others overstates the market and the post facto downgrades of previous months that correct the over-optimistic position, go unreported.
Sign in to Reply
Cowan LRA Model
6/4/2011 8:22 AM EDT
Hi Peter - I thought that I should point out a needed correction that is required to your original posting as highlighted in the following sentence abstracted from your article:
"The net result has been to strip about 91.1 million in value out of the first quarter's sales figures, which is not (should be NOW) listed by WSTS at $75.675 billion, compared with the previous figure of $75.766 billion."
Namely the "t" in "not" should be replaced by a "w" thus becoming "NOW" and therefore enhancing the intent of what you really wanted to say.
Regards, Mike C.
Sign in to Reply
peter.clarke
6/5/2011 9:33 AM EDT
Thanks Mike
Sign in to Reply
PremierSnow
6/6/2011 7:44 AM EDT
It's nothing new, they retroactively change monthly estimates all the time.
Sign in to Reply
Philosopher2
10/10/2012 2:05 PM EDT
Frankly this is sloppy journalism.
If you are going to quote numbers - what is the estimated accuracy of the Strategy Analytics data? Do they give an estimated margin of error.
0.2% could be a rounding error – either way.
Let’s wait to see some more quarters before drawing conclusions one way or the other.
It reminds me of this story http://www.eetimes.com/electronics-news/4216586/WSTS-strips-value-from-Q1-chip-market which reported a whopping reduction of the WSTS Q1 2011 estimates to $75.675 billion, down from $75.766 billion – again with no apparent understanding of the likely accuracy and margin of error of the data. Both are $75.7 billion and even one decimal point is probably below the likely margin of error in the estimate.
Sign in to Reply