Software Piracy: Evidence from Windows 7

By Susan Athey

Posted on April 3, 2014


How do consumers most commonly pirate software? Does software piracy vary across countries and intellectual property regimes? And can enforcement actions against popular pirating websites make a difference? Susan Athey and Scott Stern look at evidence from Microsoft’s Windows 7 operating system.


In the summer of 2009, Microsoft planned to release a new version of its flagship operating system, Windows 7. The redesign of the core operating system, as well as the development of bundled applications and features, represented a significant investment, with approximately 2,500 developers, testers and program managers engaged on the project for several years.

Perhaps more than any other Microsoft product before it, Windows 7 was designed with a global market in mind. The company explicitly included a large number of features aimed at serving this global market, including putting a ‘multilingual user interface’ in the most advanced version, Windows Ultimate, and creating a low-priced version, Windows Home Basic, which was targeted specifically at emerging markets.

But just weeks after the release of the final version of the software and individualized product ‘keys’, a number of websites reported that one of the original equipment manufacturer master product keys had been hacked and released onto the internet. Over the course of 2009, a number of additional leaked methods for pirating Windows 7 appeared, and by 2012, there was a large number of country-specific unauthorized Windows installation web pages, often tailored to specific languages or countries.

In a recent study, we use anonymized telemetry data from Microsoft’s experience with Windows 7 to characterize the ways in which software piracy occurs; the relative incidence of piracy across different economic, institutional and technology environments; and the impact of enforcement efforts on consumers’ choices about whether to install pirated versus paid software.

We document a range of findings. First, we document the importance of a distinctive form of software piracy: the global reuse of individual product keys, with sophisticated and active user communities that develop easy-to-follow instructions and protocols.

While the use of peer-to-peer networking sites has been associated for more than a decade with piracy of ‘smaller’ products, such as music or video, there is now a relatively direct way that any broadband user can access a fully functional version of Windows for free through the Internet.

Second, we document that the vast majority of ‘retail piracy’ can be attributed to a small number of widely distributed ‘hacks’ that are available through the internet. In addition, software piracy primarily focuses on Windows Ultimate, and piracy for this product is so pervasive that over 70% of the installations of this product worldwide are pirated.

This complicates potential efforts to deter piracy by offering ‘low-end’ versions of Windows at a greatly reduced price or free ad-supported versions. The best versions of Windows are used by pirates and they would be unlikely to choose an inferior product at the same price (zero).

Third, we evaluate how software piracy varies across different economic, institutional and technology environments. In addition to traditional economic measures such as GDP per capita (and more nuanced measures, such as the level of income inequality), we also gather data characterizing the overall quality of the institutional environment, the ability of individuals within a country to take advantage of broadband and the ‘innovation orientation’ of a country.

Our results suggest that the level of piracy in a country is closely associated with its institutional and infrastructure environment. For example, piracy is negatively associated with a number of traditional measures of the strength of legal and political institutions; it rises with greater accessibility and speed of broadband connections (faster broadband reduces the time required for pirating); and it declines with the higher innovation intensity of an economy.

Most importantly, after controlling for a small number of measures for institutional quality and broadband infrastructure, it turns out that the most natural candidate as the main driver of piracy - GDP per capita – has no significant impact on the observed piracy rate. In other words, while the pair-wise correlation between piracy and GDP per capita is strongly negative, there is no direct effect from GDP per capita.

Poorer countries tend to have weaker institutional environments - and it seems that it is the environment rather than income per se that is correlated with the observed level of piracy. Importantly, this finding stands in contrast to previous research, which has not effectively disentangled the role of institutions from the role of income per se.

Finally, we take advantage of time-series variation in our data to investigate directly the impact of the most notable anti-piracy enforcement efforts on the rate of Windows 7 piracy. Specifically, during the course of our 2011 and 2012 sample period, a number of individual countries imposed bans on Pirate Bay, the single largest source of pirated digital media on the Internet.

Though such policy interventions are endogenous (the bans arise in response to broad concerns about piracy), the precise timing of the intervention is reasonably independent of the specific piracy of Windows 7. Hence, it is instructive to examine how a change in the level of enforcement against piracy affects the rate of Windows 7 software piracy. Over a range of different anti-piracy enforcement efforts, we find no evidence for the impact of enforcement efforts on observed piracy rates.

Overall, our research offers the first large-scale observational study of software piracy. The analysis highlights the value of emerging forms of passively created data, such as the Windows telemetry data, and also the role of both institutions and infrastructure in shaping the total level of piracy.

This article summarizes ‘The Nature and Incidence of Software Piracy: Evidence from Windows’ by Susan Athey and Scott Stern.

Susan Athey is a Professor of Economics at Stanford Graduate School of Business and a member of the Toulouse Network for Information Technology (TNIT). Professor Scott Stern is the Chair of the Technological Innovation, Entrepreneurship, and Strategic Management Group at the MIT Sloan School of Management.

The preceding post is republished on TAP with permission by the Toulouse Network for Information Technology (TNIT). “Software Piracy: Evidence from Windows 7” was originally published in TNIT’s March newsletter.