The Economic Question
How does the market price companies exposed to similar growth, profitability, capital intensity and risk? A multiple is useful only when the numerator and denominator describe the same economic perimeter and period.First-Principles Derivation
Peer enterprise value = peer multiple × target operating denominator
Peer equity value = peer enterprise value − target net debt − senior obligations + non-operating assets
Peer value per share = peer equity value / diluted equivalent shares outstanding
Currency cancels inside a properly constructed ratio, because numerator and denominator use the same currency. It does not disappear when the implied enterprise value is translated into a target share price. That final bridge still requires source-based exchange rates and a security-specific share denominator.
Peer Membership Is Preserved, Not Curated
Parallax begins with a company-specific peer group, assembled from the issuer’s own economic classification and size band rather than from an analyst’s discretion, and preserves its members. The output lists them, so a reader who disagrees with the membership can say which company is wrong and why. An industry label does not grant permission to remove a company from that group. An extreme observation can be flagged, explained and tested without deleting the issuer.
The Statistical Center
Each eligible issuer-level observation is translated into an implied value for the target. The displayed range is built from the distribution of those observations, using quartiles where the sample supports them and observed extremes where it does not. The general peer-evidence range keeps the sample median as its contextual midpoint. The client price-target workbook is a separate calculation and uses a robust quartile-weighted center instead. Both draw on the same eligible observations, and each labels the statistic it applies rather than substituting one center for another without saying so. Outliers remain in the evidence and are flagged using relative and distributional diagnostics. They are not silently deleted or winsorized. Parallax also calculates leave-one-out and unflagged centers as stability diagnostics. Those diagnostics challenge the primary center. They do not quietly replace it.Why a robust center earns its place
Consider eight eligible peer observations implying share values of 2.7, 24.0, 39.3, 40.5, 46.7, 53.8, 67.8 and 84.2 in local currency. The arithmetic mean is 44.9 and the median is 43.6. The first and third quartiles are 35.5 and 57.3. Tukey’s trimean weights the median twice and each quartile once:Trimean = (Q1 + 2 × median + Q3) / 4 = (35.5 + 2 × 43.6 + 57.3) / 4 = 45.0
The result does not let the 2.7 low observation or the 84.2 high observation dominate, yet it recognizes that the upper half of the distribution is wider than the lower middle. The endpoints remain visible. Robustness comes from the statistic, not from quietly deleting awkward peers.
Figures are displayed to one decimal place. Calculations retain full precision.
When Two Metrics Disagree
When two approved metrics both produce full ranges and their centers are close, the ranges can be combined component by component. When they disagree materially, the primary metric leads and the second remains visible as contested evidence. Divergence is information. It is not an inconvenience to average away.A Real-World Edge Case
A narrative-driven company can trade far beyond ordinary multiples for its sector, sometimes anchored by enthusiasm around a private competitor. Removing that public company because its valuation is extreme would turn the peer process into an opinion filter. Allowing the extreme observation to define the center would be equally weak. The disciplined treatment preserves the company, shows its raw implied value, flags its influence, and compares the all-observation center with leave-one-out and unflagged diagnostics. A client can then see both the market’s enthusiasm and the range that is less sensitive to a single story stock.Interpretation Note
A lower peer multiple may reflect lower growth, lower ROIC, weaker balance-sheet quality or greater risk. Read P/B alongside ROE and cost of equity; P/E alongside growth, payout and risk; EV/EBIT alongside operating growth and reinvestment; and EV/Revenue alongside future margins and capital intensity.
How to Assess the Result
- Confirm that target and peers use the same period and denominator basis.
- Inspect every member, including cross-listing deduplication and constructed additions.
- Compare the primary center with leave-one-out and unflagged diagnostics.
- Translate the multiple back into an implied margin, ROE or growth assumption.
- Check the enterprise-to-equity bridge and the diluted share count.
When to Reassess
The peer set does not share the relevant economics; the target denominator is nonpositive or inconsistent; too few same-basis observations survive; a cross-listing appears twice as separate companies; or the implied value depends on an exchange rate or share count that does not reconcile.How Parallax Applies This
Parallax exposes the member list, metric basis, issuer-level observations, outlier flags, statistical center, constructed-peer count, and the translation to value per share. Peer context never rescues a blocked intrinsic valuation, and it is never called intrinsic value.Archetype Conventions
Which multiple family a company’s balance sheet actually justifies.
Price Targets and Ratings
How peer evidence and consensus combine into a client-facing target.