From Pixels to Provenance: Empirical Study of How NFT Valuation Evolves from Traits to Reputation

Why do some NFTs sell at multi-fold premiums while otherwise similar tokens trade at discounts? Existing research on NFT pricing has documented average effects of intrinsic traits and wallet histories, but evidence on heterogeneous pricing mechanisms and on trading patterns associated with the observed premiums remains scarce. We address this gap using transaction-level data on 48,319 sales of the Bored Ape Yacht Club collection (2021–2025) and applying causal forests with honest splitting and orthogonalization. First, we observe that pricing follows a temporal valuation hierarchy: traits dominate first-sale prices, while wallet-level signals explain treatment-effect heterogeneity in secondary markets, with observed covariates accounting for 59–88% across wallet-characteristic specifications. Second, lagged wallet effects dominate contemporaneous ones: in intermediary-mediated transfers, high prior-seller wallet value is associated with a 42% reduction in next-sale price growth, while high prior-seller gas expenditure is associated with an 85% premium. Third, joint treatments combining low-value buyer-seller wallets with high gas-fee activity are associated with premiums of 240–726% in narrow subsets of intermediary-mediated transfers (largest estimate based on n=105); these magnitudes are consistent with reputation signaling and coordinated activity rather than independent trading, but should be interpreted cautiously given the small samples. These findings extend singular-goods theory to on-chain markets, document pricing patterns consistent with coordinated trading, and inform marketplace design and oversight of digital-asset markets.