On the Use of the Bass Model for Forecasting Pecuniary Damages: a Reappraisal

In a 2006 paper in the Journal of Forensic Economics, Tomlin & Wazzin purport to show the inapplicability of the Bass Model for routine, mundane estimation of pecuniary damages (Tomlin & Wazzan, 2006). We agree. A Bass model is better suited for appraising but-for estimates of lost sales when the environment constitutes a homogeneous product viewed as innovative or novel by its prospective customers and sales and marketing efforts benefit from diffusion via social networks. We argue that when confronted with an underlying diffusion data generating process of a but-for sales effort, the task at hand is twofold: (i) to determine the rate of sales increase, and (ii) to identify the apex of the but-for sales path. Given these tasks we show that a linear model is unsuited for purposes of illustrating counterfactuals. The Bass model, on the other hand, reproduces the underlying data-generating process more adequately. We re-examine the Bass model using a more conventional simulation study to compare the accuracy of the Bass Model to a competing linear model. Our results uphold the generality of the Bass model – especially when modeling counterfactual performance of products perceived as novel and innovative by its prospective customers.