Playbook › Playbook
Correcting one name's forward multiple while leaving its peers on vendor data invents or destroys a discount
Claim
pitfall-vendor-forward-eps-is-the-wrong-fiscal-year and
pitfall-vendor-forward-eps-is-stale-on-the-day-of-a-guidance-cut both catch a single name's
broken forward multiple. This note covers what happens next — and it is the more dangerous
error, because it arrives disguised as diligence.
Once you have corrected the target, you must correct every peer before comparing. A half-corrected table is worse than an uncorrected one: the uncorrected table is consistently wrong and the relative signal survives; the half-corrected table is inconsistently wrong and the relative signal is fabricated.
The reason peers are almost always wrong too: vendor forward EPS breaks for structural reasons — a consensus not refreshed after a guidance reset, a wrong fiscal year, a stale pre-spin basis. Those causes are correlated across a sector, because sector peers report on similar calendars, re-guide in the same weeks, and are covered by the same feed logic. The default assumption should be that if the target's forward is broken, the comps are broken too.
Evidence — SPGI, 2026-08-12
Yahoo showed SPGI at PE(fwd) 20.18 and MCO at 25.20, implying a 5.0-turn / ~20% discount
for SPGI against its closest comparable — the other half of the ratings duopoly.
SPGI's forward was genuinely broken: Yahoo's implied forward EPS of $20.24 was a pre-spin figure (the Mobility spin completed 2026-07-01), while the actual post-spin FY26 guide was $17.50–17.75. Correcting only SPGI:
| Price | Vendor fwd P/E | Vendor implied EPS | Verified FY26 guide | True fwd P/E | |
|---|---|---|---|---|---|
| SPGI | 408.39 | 20.18x | $20.24 | $17.50–17.75 | 23.2x |
| MCO | 477.24 | 25.20x | $18.91 | $16.50–17.00 | 28.5x |
| FDS | 277.04 | 14.10x | $19.65 | $17.25–17.75 | 15.8x |
Every one of the three was inflated — SPGI by 14.8%, MCO by 12.9%, FDS by 12.3%. The cause differed by name (SPGI: stale pre-spin basis; MCO and FDS: consensus not refreshed after mid-2026 guidance raises), but the direction was identical.
The consequence, in the order it actually happened during the analysis:
- Manager corrected SPGI 20.2x → 23.2x, then inferred the MCO discount collapsed from 5.0 turns to ~2 turns, and passed that to the Valuation analyst as an established fact.
- Valuation built its peer table, SOTP Ratings multiple, fair-value range and verdict on the 2-turn figure, and wrote that the "identical franchise, unjustified gap' framing does not survive the correction... this materially weakens the cheapness case."
- Fundamentals independently pulled MCO's actual guide and found 28.5x, restoring the discount to 5.3 turns / −18.6% — marginally wider than the uncorrected vendor figure.
The half-correction produced a worse answer than doing nothing. The uncorrected vendor table said 5.0 turns; the truth was 5.3 turns; the half-corrected table said 2.0 turns.
How to apply
- Correct the whole row or none of it. If you fix the target's forward EPS, pull the guided EPS for every comp in the table before computing a single relative number.
- Prefer guided EPS over consensus for the whole table. Company guidance is a primary source with a known basis; vendor consensus is an aggregation with an unknown one.
- State the basis in the table itself — a column showing each name's EPS figure and where it came from makes the error impossible to hide.
- Treat a vendor forward-EPS error as field-wide until proven name-specific. One broken forward in a sector is a reason to check the others, not a reason to feel good about the target.
- Watch for this specifically after any event that resets a basis across a sector at once: a spin, a restatement, an accounting-standard change, or a common guidance season.
What would falsify this
Finding a sector where vendor forward-EPS errors are genuinely idiosyncratic — i.e. the target is broken and the peers verify clean against their own guidance. That would make the half-corrected table accidentally right, though it would still be right for no reason.
History
- 2026-08-12 — created from the SPGI
/analyze, where the error was made by the manager, propagated into a dispatched Valuation brief, and caught by the Fundamentals analyst before it reached the verdict. Related: [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]], [[pitfall-vendor-forward-eps-is-stale-on-the-day-of-a-guidance-cut]], [[principle-primary-source-beats-vendor]].