Cowan LRA model: Overview plus latest global semicon sales forecast numbers

This is a continuation of my coverage of the fortunes of the global semiconductor industry. I would like to acknowledge and thank Mike Cowan, an independent semiconductor analyst and developer of the Cowan LRA model, who has provided me the latest numbers.

August 2010′s ‘actual’ global semiconductor sales numbers are scheduled to be released by the WSTS, namely the August HBR (Historical Billings Report) on or about October 4.

In anticipation of the WSTS release Cowan has shared an analysis feature of the Cowan LRA Model for forecasting worldwide semiconductor sales; namely, the ability to provide a ‘look ahead’ scenario analysis for 2010′s global semiconductor sales forecast as a function of next month’s (in this case August’s) actual global semi sales estimate.

The specifics of the scenario analysis are presented in the following paragraphs and detailed in the scenario analysis matrix table provided here.Source: Cowan's LRA model.

In order to demonstrate this capability, Cowan has selected a range in possible August 2010 sales; in this particular scenario analysis, a sales range from $23.95 billion to $26.95 billion in increments of $0.5 billion was chosen as listed in the first column of the table.

This estimated range of actual sales is ‘centered around’ the actual August sales forecast estimate of $25.448 billion as determined by last month’s (July) run of the model. The corresponding August 3MMA sales forecast estimate that the model put forth is $25.723 billion.

The overall year 2010 sales forecast estimate for each assumed estimated August sales number over the selected range of August actual sales estimates is calculated by the model, and is shown in the second column of the table.

The third column reveals the resulting yr-o-yr sales growth estimates compared to year 2009′s actual sales (of $226.3 billion).

The fourth and fifth columns show the corresponding three Month Moving Average (3MMA) sales estimate and the associated year-on-year sales growth relative to August 2009′s 3MMA sales (of $19.381 billion), respectively.

Finally, the sixth column lists the associated Momentum Indicator (MI), which is defined and discussed below.

July 2010’s actual semiconductor sales (of $24.568 billion) came in higher (by $1.180 billion) than the model’s last month’s July 2010 sales forecast estimate (of $23.388 billion) representing a plus 5 percent delta comparing July 2010′s actual sales number (published by the WSTS) to the projected forecast estimate ‘put forth’ by the Cowan LRA forecasting model and reported last month. This percent delta represents the Cowan LRA Model’s MI.

The MI is defined as the percent difference between the actual sales for a given month — in this case July 2010’s just published actual global sales of $24.568 billion and the forecasted sales estimate for July 2010, that is, $23.388 billion, which was calculated and published last month.

The MI can be either positive or negative and is a measure of the percent deviation of the actual monthly sales number from the previous month’s prediction derived by the model’s linear regression analysis of the past 26 years of historical, actual monthly global “sales experience” as gathered and published, each month, by the WSTS.

Note: August 2010’s sales forecast estimate is projected to be $25.448 billion.

Therefore, the table above reveals that depending on the actual WSTS to-be-released August 2010 global semiconductor sales number, the forecasted 2010 sales estimate, as determined by the model, could vary between $303.19 billion and $307.90 billion, while the corresponding 2009 –> 2010 sales growth estimate could vary between 34 percent and 36 percent, respectively.

Also note that the previously published Cowan LRA Model’s 2010 sales forecast update, which was based upon July 2010′s actual sales (of $24.568 billion), came in at $305.729 billion, corresponding to a sales growth forecast estimate of 35.1 percent as summarized earlier.

Using this analysis capability, the model provides a ‘sensitivity output’ of the ‘expected’ 2010 sales forecast (and year-on-year forecasted sales growth) as a function of the yet-to-be-published actual August sales number.

Therefore, employing the table, one can a-priori “select” an August sales number (in the range shown) and immediately see what the model would predict for a 2010 sales forecast estimate along with its corresponding sales growth expectation in advance of the actual to-be-published final August sales result.

Stay tuned for the WSTS to publish (expected on Oct. 4, 2010) the actual August 2010 sales numbers. One can then ascertain the model’s latest forecast outlook as abstracted (or extrapolated) from the table even before Cowan runs his model and subsequently publishes the updated forecast numbers based upon August’s actual sales results.