Measuring the usage effects of tying a messenger to Windows: a treatment effect approach

Abstract

The US case on tying Microsoft Internet Explorer to Windows has received much attention. In Europe, a similar case of tying the Microsoft media player to Windows appeared. Recently in Korea, another similar case of tying a Microsoft messenger to Windows occurred. In the messenger tying case (as well as in the other tying cases), Microsoft's main defence seems to be threefold: tying enhances efficiency, the Microsoft product is better or better marketed and tying is inconsequential because the user can easily download free competing products. The paper empirically addresses the third point. Korean data, used as evidence in the trial of the case, reveal that tying the Microsoft messenger to Windows increased the probability of choosing the Microsoft messenger as the main messenger by 22% for Windows Millennium and 35% for Windows XP. There is also evidence that tying shortened the duration until the Microsoft messenger is adopted by about 2-4 months, compared with the duration until the adoption of a competing messenger. Hence tying provided Microsoft with an almost instant non-trivial advantage in the messenger market 'race'-the advantage derived from the dominant position in the operating system market.