Variable Marginal Propensities to Pirate and the Diffusion of Computer Software

In this paper, we empirically investigate the dynamics of the marginal propensity to pirate for computer software. We introduce a state space formulation that allows us to estimate error structures and parameter significance, in contrast to previous work. For data from 1987-92, we find a rising propensity to pirate as the number of existing pirate copies increases, and higher late piracy incidence than implied by static models. We strengthen prior results on the impact of piracy in the spreadsheet market, finding it to be the only significant internal influence on diffusion. However, when we allow for negative error correlation between legal and pirate acquisitions, we contradict earlier work by finding that, in the word processor market, piracy did not contribute to diffusion and only eroded legal sales.