Latest Cowan LRA Model's 2013 And 2014 Global Semiconductor Sales and Sales Growth Expectations
(Derived From the WSTS's Recently Posted May 2013 Historical Billings Report, HBR)
July 5th, 2013
I again would like to share with my weSRCH readers the just updated 2013 and 2014 forecast numbers derived by
exercising the Cowan LRA, Linear Regression Analysis, forecasting model on May 2013’s actual sales ($24.343 billion)
as gleamed from the WSTS’s May 2013 HBR, Historical Billings Report, which was recently posted (7-05-13) on their
The first table shown to the left provides a high-level summary overview of the model’s latest updated sales and yearover-year sales growth forecast expectations for both 2013 and 2014 shown on a quarterly, half-year, and full year basis.
The second table summarizes the latest, updated "thinking" of a number of other prognosticators (including the WSTS’s
latest, published last month, sales growth forecast number) for 2013 sales and sales growth numbers -- ranked from high
to low on a year-over-year sales growth expectation basis.
Note that the latest Cowan LRA Model's sales growth expectation for 2013, derived from the WSTS's just released May
2013 HBR actual sales result, improved by 1.0 percentage point to 4.0 percent – which is higher than the model’s previous
month’s 2013 sales growth prediction of 3.0 percent.
Additionally, with the model extended to "pick up" the four quarters of 2014 thus provides the model's updated forecast
estimates for next year's quarterly sales and corresponding sales growth forecast estimates. Consequently, the resulting full
year 2014 sales growth expectation relative to the model's just updated full year 2013 sales forecast estimate of $303.197
billion came in at 5.7 percent which is a slight downward drop of 0.1 percentage points relative to last month’s 2014 sales
growth forecast of 5.8 percent.
The third table displays the monthly tracking of the evolution of the model's 2013 month-by-month sales and sales growth
“thinking" for the past eleven months as well as over the past three months for 2014.
Moreover, one must be reminded that these monthly sales forecast estimates are derived from a strictly mathematical
model that depends on linear regression analysis techniques operating on the past 29 years of historical, actual sales
numbers from 1984 through 2012 that are gathered and published by the WSTS each month (see next page for model
Finally in the 3rd table, I would like to draw the reader’s attention to the YTD 2013 Actual Sales Grwth TREND (%) line
which reveals the month-to-month evolution of the actual cumulative sales number compared to the corresponding month
of last year’s YTD sales. As shown this trend has exhibited a turning point (point of inflection) this month in concert with
the upward trend in the latest 2013 global semiconductor sales growth expectation as put forth by the model.
The Cowan LRA Model’s Semiconductor Forecast – A Succinct Overview of Its Forecasting Methodology
Developed by Mike Cowan (Independent semiconductor industry analyst and market watcher)
A unique and practical global semiconductor sales forecasting model developed by the author facilitates the determination
of future global sales of the semiconductor industry. The Cowan LRA (Linear Regression Analysis) Model, which forecasts
global semiconductor sales, is a mathematically based model that features statistical analysis of the past 29 years of
historical, monthly actual global semiconductor sales that are collected and published by the World Semiconductor Trade
Statistics (WSTS) organization. It is a dynamic, mathematically pure view of near-term worldwide semiconductor sales
looking forward over the next five quarters.
Since the model is purely mathematically-based, it is devoid of any economic assumptions or emotional biases. It
exploits linear regression analysis operating on the "appropriately transformed" actual monthly semiconductor sales
numbers that are gathered and published by the WSTS. This mathematical transformation of the actual monthly sales
numbers thereby "renders" the global semiconductor sales data highly linear and, therefore, very amenable to linear
regression statistical analysis techniques. The numerical transformation of the past 29 years of monthly actual sales
numbers -- from 1984 through 2012 -- that is invoked is not a complicated mathematical expression but very straight
forward and "makes sense physically," thereby yielding extremely high linear regression correlation coefficients of greater
than 0.97. In exercising this forecasting model each month, a total of eight distinct sets of linear regression parameters (of
the format: y = mx + b) are employed to calculate the resulting global semiconductor sales forecast estimates for each
one of the next seven quarters associated with the model's forecast horizon as well as next month’s sales forecast
It is emphasized that each month's actual global sales number released by the WSTS (as posted on its website) is a
"lagging indicator" since it is released a full month after the fact, that is, it is backward looking. The Cowan LRA Model,
however, "turns" this lagging monthly sales number into a "leading indicator" by virtue of its near-term forecasting
capability looking out over each of the next seven quarters. This is the "beauty" of the model and, therefore, makes it
dynamic in the sense that it is run each month using the most recent actual global semiconductor sales numbers
published by WSTS via their monthly Historical Billings Report. Thus, it can rigorously "track" the near-term sales forecast
outlook of the global semiconductor industry on an almost real-time basis. Consequently, the model's monthly sales
forecast expectations do not "sit still" but evolve each month since conditions change rapidly and unexpectedly in the
semiconductor industry, and market forecasters are hard pressed to keep up with these monthly changes. How can
industry management be sure that a forecast issued two, three or more months ago is still relevant to what's happening in
today's dynamic semiconductor market? With the Cowan LRA Model’s forecast results being issued each month, this can,
therefore, help eliminate this very real dilemma.
ABOUT THE AUTHOR
Mike Cowan, the developer of the Cowan LRA Model, is a 48-year semiconductor industry veteran. He enjoyed a 36.5-year career
at IBM's Microelectronics Division in East Fishkill, N.Y., where he was involved in many facets of semiconductor development
and manufacturing engineering, including both technical and management responsibilities. Over his last ten years at IBM, as a
senior technical staff member, he was involved in strategy development and competitive analysis activities focused on the
semiconductor industry, and developed a wide range of top-down and bottom-up models to predict the dynamics of the
semiconductor industry. After retiring from IBM in Jan. 2002 he became an independent semiconductor industry watcher /
analyst, upgraded and perfected his forecasting model and then launched his monthly global semiconductor sales forecast
activities in August 2002. Cowan earned both BS and MS degrees in physics from Wayne State University in Detroit, Michigan,
and an MS in electrical engineering from Syracuse University in New York. For more information and / or questions regarding
either the model or the results, readers are encouraged to contact the author directly at mikedcowan(at)verizon(dot)net.