This paper presents the first empirical survival analysis of Brazilian registered trademarks disaggregated by all 45 Nice International Classification classes. Using right-censored survival data drawn from the Brazilian National Institute of Industrial Property (INPI) registry — covering 1,555,183 unique trademark processes with publication dates spanning 1975 to 2020 — we calibrate Weibull proportional hazard parameters (k and λ) for each Nice class via Maximum Likelihood Estimation (MLE). Results show that trademark mortality follows a Weibull distribution with shape parameter k > 1 uniformly across all 45 classes, indicating monotonically increasing hazard rates — trademarks face rising extinction as they age. Characteristic lifetime (λ) ranges from 44 to 75 years, and the 10-year survival probability (S(10a)) ranges from 0.935 to 0.957. The systematic underestimation of extinction risk by the Gordon Growth Model's implicit assumption of perpetual survival is quantified for each Nice class, yielding per-class Gordon overvaluation biases between 4.3% and 6.5% at a 10-year horizon. These estimates provide actuarially grounded discount factors for trademark valuation under the Relief from Royalty method, with direct application in Purchase Price Allocation (IFRS 3 / ASC 805), brand impairment testing (IAS 36 / ASC 350), and corporate litigation appraisals. While Artificial Intelligence was employed for linguistic translation and English-language polishing, the entire theoretical framework, mathematical derivations, empirical calibrations, and the final manuscript composition were exclusively developed by the autho
Ruben Freitas (Sun,) studied this question.