The Nature and Incidence of Software Piracy: Evidence from Windows

Copyright and Trademark and Intellectual Property

Article Snapshot


Susan Athey and Scott Stern


National Bureau of Economic Research Working Paper #19755, December 2013


This study shows that most pirated copies of Windows 7 come from a few leaked product keys. Newly released software is more likely to be pirated. Respect for law and property within a region is associated with lower piracy rates.

Policy Relevance

Enforcement efforts do not affect piracy rates. Offering a low-cost version of a product is unlikely to reduce piracy.

Main Points

  • When Windows 7 was released, Microsoft offered a low-priced version for emerging markets.
  • Software makers often allow users of pirated products to use and update their products, because these users are often exposed to malware, which creates problems for all users.
  • Policies intended to address piracy vary depending on perceptions of its origins.
    • The case for aggressive enforcement is strong if piracy stems from lack of enforcement.
    • The case for tolerance is strong if low-income consumers cannot afford the product.
  • This study used data generated during Windows automatic updates to study piracy.
    • The data included the user’s product key, the model and make of the user’s machine, and the user’s location.
    • The average piracy rate in each of the 95 countries studied is 40%.
  • Only 12 leaked product keys were responsible for 90% of piracy of new releases of Windows.
    • Offering low-priced “bare-bones” versions of software will not reduce piracy, because a full-featured version will be available for free.
    • More than 70% of copies of Windows Ultimate were pirated.
  • Some poor countries lack quality institutions to support property rights and the rule of law, and this is correlated with high piracy rates.
    • Johannesburg and Beijing have about the same GDP, but piracy rates in Beijing are more than twice as high.
    • Moscow’s piracy rate is four times as high as that of Sydney, but Moscow’s GDP is not much lower.
    • The rate of GDP per capita does not affect piracy rates.
  • Enforcement actions such as banning the Pirate Bay website do not affect piracy rates.
  • Software piracy is a meaningful economic phenomena; policymakers in the United States may understate the problem because piracy rates in the United States are low.


Get The Article

Find the full article online

Search for Full Article