Valuing Businesses Using Regression Analysis

CHAPTER 6Polynomial Regressions

In probably one out of twenty regressions performed on samples of comparables, the linear regression methodologies reviewed in the previous chapters will appear to produce unreasonable results. This occurs most frequently when the subject company itself is an outlier. That is, the subject's operating profit margin (SDE%) was well above or was below most of, or all of, the transactions in the sample. This is not to say that medians will be the better choice of metrics here because medians fail under these circumstances as well. Exhibit 6.1 is a classic example. To illustrate, the subject's SDE% in this example is 21.5%, well above almost all of the transactions in the sample.

The intersection on the regression line at an SDE% of 21.5% suggests that a cash flow multiplier of approximately 3.25 is appropriate. However, we observe that there are five highly profitable transactions in the sample that have multipliers between 3.4 and 3.6. Clearly the linear regression appears to undervalue the subject as the regression line was not responsive to those highly profitable transactions.

The same situation appears at the low‐profit end of the SDE% spectrum. If our subject's SDE% was in the 2% to 4% range, the regression line does not seem to recognize the fact that two comparables in that level of profitability produced multipliers in the 4.4 to 4.5 range, well above the regression line.

EXHIBIT 6.1 Linear Regressions on Subject's That Are Outliers