How Well Do Comparable Transaction Filters Predict Biotech Valuations? Evidence from 1,976 Financing Events

<div> Venture capital investors and investment banks routinely construct comparable transaction sets by filtering on clinical development stage, therapeutic indication, and drug modality to value biotech companies. Despite the near-universal adoption of this framework, limited empirical evidence supports the assumption that these variables meaningfully predict pre-money valuations. I test this assumption using 1,976 biotech financing transactions spanning 2013 to 2026, sourced from SEC EDGAR filings and cross-referenced with ClinicalTrials.gov for independent clinical taxonomy assignment. I decompose valuation variance using OLS with robust standard errors and Ridge regression with ten-fold cross-validation. The full model — incorporating clinical stage, therapeutic indication, drug modality, deal type, and calendar year — explains approximately 8.8% of log pre-money valuation variance out of sample. The partial ????2 decomposition contradicts conventional practice: deal type contributes 4.2 percentage points, therapeutic indication 1.4, clinical stage 1.2, and drug modality 0.2. Stage — widely treated as the primary valuation driver — ranks third. The stage ladder is largely flat between Phase 1 and Phase 2/3; only Phase 3 companies are reliably distinguishable from Phase 2 firms. The strongest individual effect is a rare disease premium of approximately 3.5 times. When the sample is restricted to IPOs — the transaction type where comparable analyses are most commonly employed — the model’s out-of-sample ????2 turns negative, indicating zero predictive power. These findings suggest that while the comparable transaction framework may serve governance and anchoring functions, it has limited empirical support as a price discovery mechanism for biotech financings. </div>