The 17.4x trailing multiple sits well below the SPGI (26.9x), MSCI (34.6x) and TRI (28.6x) comparables, which suggests the market is discounting FactSet's structurally lower gross margin — 54.0% versus Thomson Reuters' 75.8% — and its sensitivity to buy-side headcount. The 1.73 PEG reads as reasonable only against an assumption of mid-to-high single-digit growth; with diluted EPS declining in each quarter of 2026 despite rising revenue, that assumption appears more fragile than the multiple implies, so the discount may reflect earnings quality concerns rather than a permanent valuation gap.

DuPont decomposition indicates the decline is not a leverage artifact — equity-to-assets has held broadly stable between 0.47 and 0.51 — so the compression appears driven by net income contraction rather than by equity release through buybacks, which would normally flatter ROE mechanically. With fee income accounting for effectively 100% of revenue, the ROA path to 2.9% reflects a margin problem rather than a spread or credit problem, and the persistently negative net interest income line adds a small but recurring drag on reported returns.

An equity-to-assets ratio of 0.48 alongside debt-to-equity of 0.71 is adequate in absolute terms but shows no visible expansionary headroom: with 2026 operating cash flow of $822.2M nearly equalled by $808M of dividends and buybacks, deleveraging capacity would appear to require moderating capital return, which has historically been management's primary shareholder-facing commitment. Investors should monitor whether repurchase pace is reduced before assuming the balance sheet can absorb further acquisitions on top of the CUSIP Global Services financing load.

Fee income of 100.0% to 106.8% of revenue across the series confirms the business is entirely subscription-driven, so the negative NIM print appears to reflect financing costs on the post-CGS debt load rather than any lending activity — a structural drag, not a cyclical one. The upward drift in the cost ratio alongside declining EPS suggests royalty payments and technology spend are rising faster than pricing, which would be consistent with structural margin compression rather than the operating leverage management has historically guided toward.

The size of the provision line — approximately 2.8x reported net income in 2026Q4 — is inconsistent with an asset-light model that discloses essentially no loan book, so the figure may reflect reclassification or non-operating charges rather than genuine credit exposure. Until the composition of the line is clarified, it should not be read as evidence of deteriorating asset quality, nor should it be excluded from earnings without further disclosure; both interpretations carry valuation consequences.