The Contaminated Comparable Set
A valuation analyst is preparing a fair value measurement of a minority (10%) non-controlling equity interest in a privately held specialty-chemicals manufacturer. Management has supplied a list of 8 'comparable' acquisitions from the last four years. Reviewing the list, the analyst finds: 3 transactions were control acquisitions with synergies explicitly quantified in the press releases; 2 were distressed sales during a regional credit crunch; 1 was a carve-out of a single product line, not a whole company; and 2 are genuine comparable minority trades, though one occurred 41 months ago. The engagement letter says the report must comply with a recognized valuation standard requiring the analyst to select and document the most relevant market evidence. The client's CFO emails: 'Just average all eight multiples so we don't get second-guessed by the auditors.'
1. How should the analyst respond to the CFO, and what is the defensible treatment of the eight data points?
Read the worked answer
The analyst should decline to mechanically average all eight and should explain why in writing.
Reasoning:
- A multiple is only meaningful when the numerator (price) and denominator (earnings) reflect the same bundle of rights and risk. Control premiums and synergies are embedded in the three control prices, so applying those multiples to a 10% minority interest overstates value.
- Distressed-sale multiples reflect seller duress, not fair value in an orderly transaction. The standard's definition of fair value presumes an orderly transaction between market participants at the measurement date, so distressed evidence is generally excluded or heavily discounted.
- The carve-out multiple prices a product line, not an enterprise; its denominator is not comparable.
- The 41-month-old trade is stale relative to the measurement date and to industry conditions.
Defensible treatment: give primary weight to the two genuine minority trades, adjust for staleness, and disclose the excluded data with reasons. If only one or two points survive, say so and corroborate with a secondary approach rather than manufacturing a false consensus. A documented, reasoned selection is what the standard requires; averaging everything is not 'conservative,' it is unsupported.
2. The client argues that using only two data points is 'too thin' and asks the analyst to widen to all eight for 'statistical robustness.' How should the analyst frame the trade-off between sample size and relevance?
Read the worked answer
More observations do not improve an estimate if the observations measure a different thing. Widening the set trades a relevance error for a precision illusion.
Framing for the client:
- Each excluded transaction differs from the subject on a dimension that moves value (control rights, duress, asset scope, time). Including them introduces bias whose direction is known (upward for control/synergy, downward for distress), not random noise that averages out.
- Statistical robustness assumes the observations are draws from the same population as the subject. Here they are not.
Practical path: keep the relevant set as primary, then corroborate with an independent method (e.g., an income approach or a guideline-company method using minority trading prices). Report a range and explain what drives it. If the client still wants breadth, the analyst can present the full set as a sensitivity labeled as such - never as the concluded value - and document that the conclusion rests on the relevant subset.