Valuing Businesses Using Regression Analysis

CHAPTER 3Identifying Outliers Using Regression

There will always be a few transactions in every sample whose actual selling price was radically different from the price suggested by the regression trend line (i.e., they are significantly out of alignment with the rest of the market.) The regression analysis not only plots a line that best represents where the market is but also calculates what is referred to as standard error lines. The standard error is a statistical measurement similar to standard deviation. The difference is the standard deviation is a measure of dispersion around a single point in a sample (the mean), whereas the standard error is the same measure dispersion around the regression line. The standard error calculates the upper and lower boundaries between which most of the multipliers (68%) in a sample should fall.

Theoretically, 16% of the sample's transactions should have revenue multipliers that fall above the upper standard error line and 16% of transactions should fall below the lower standard error line. Those transactions that fall outside these boundaries are companies whose selling‐price multiples were so far above or below the rest of the market that the transactional data must be considered flawed or not conforming to the stated standard of value. These “outliers,” as they are referred to, will be removed from our sample of comparables.

The sample from Exhibit 2.3 produced a standard error of 0.10. Exhibit 3.1, below, adds two dotted lines to the ...