Résumé
We investigate empirically the relationship between software piracy and GDP per capita by considering non-linear effects. We use a panel data analysis for 100 countries over a period of 15 years. We remedy several previous econometric and methodological shortcomings and show that piracy follows a Kuznets-like curve. Concretely, piracy first increases with the level of GDP per capita, reaches a maximum, and then decreases at higher levels of income. Making people richer can be the best way to decrease piracy over the long-term horizon. Intellectual property rights holders should not aim for a decrease of piracy per se, but rather a decrease of piracy in those circumstances where it is most likely to be substituted by legal sales. Economic growth can generate by itself incentives to curb piracy.