R-Squared and Business Comparables
Dec 01 2017
How to determine the credibility of data when using a business comparables sales database?
The R-Squared method compares a data set to the line of regression on a statistical chart. In this case, we plot PeerComps’ transaction set, against the linear regression representing the relationship between Price and EBIDTA.
The R-Squared function score ranges from 0% to 100%, where 0% represents the lack of cohesion of a data set to its regression mean, and a score of 100% represents perfect cohesion. In business valuation, the industry generally considers an R-Squared of 75% and above as reasonable.
Above is a plot of PeerComps business comparables data comprised of all Medical Equipment and Supplies Wholesalers’ transactions (NAICS code 423450) between a revenue size of $1.0 million and $10.0 million. Typically, outliers are eliminated to avoid skewing of data, but since there were no significant outliers, no data points were eliminated in this study. Also, it is worth noting that all transactions in the data set are SBA approved, therefore eliminating the most outliers as a factor.
Here is where PeerComps scored:
- 5% when comparing Price vs. EBIDTA
- 2 % when comparing Price vs. SDE
- 2% when comparing Price vs. Revenue
When speaking with a client based on PeerComps business comparables data shown above, you can provide a business valuation or price at approximately 95% accuracy based on EBIDTA or SDE. This also provides an explanation as to why a business valuation based on revenue is not recommended is most likely misleading. Why? Because, the empirical study shows that this market trades on multiples of earnings, and really cannot be gauged by the overall revenue with any certainty.
The R-Squared test is a function available in Excel (RSQ). I would challenge you to run this test of data cohesion on other databases. I have, and the clear winner is PeerComps. In other databases, I occasionally get a score in the 75% – 80% range, however, in most cases to get there I would need to eliminate numerous outliers. I have found PeerComps to be more accurate, composed of reliable data, and as with the example above I did not have to remove or add any data points.
In conclusion, PeerComps allows you to accurately present your client a valuation based on actual market transactions with a significantly higher level of accuracy. The R-Squared study also helps you explain how the market trades, and the legitimacy of the valuation method. PeerComps has a track record of accuracy that is completely unprecedented, and superior to other comparable sales databases.
By: Steve A. Mize, ASA with Douglas E. Teten
Steven A. Mize is the Managing Partner at GCF Valuation, Inc. and PeerComps, Inc. Steve is an Accredited Senior Appraiser with an extensive background in the business valuation. Please feel free to contact him with questions on this article or with any other business valuation questions.
Douglas E. Teten is a Valuation Analyst at GCF Valuation and a self-proclaimed data geek.