An increasing number of scholars have begun to argue that market forces may not automatically select the best technologies or products. During the 1980s and 1990s, Paul David and Brian Arthur published several papers in which they asserted that sub-optimal or inefficient technologies can become "locked in" as industry standards, and, in instances where there are significant network effects, these inefficiencies may persist for extended periods of time (David 1985; Arthur 1989, 1990). (1) David's best-known work in this area is his discussion of the layout of the QWERTY typewriter keyboard, and Arthur's most popular example concerns the struggle for supremacy over VCR format (David 1985; Arthur 1990). They concluded that the commercial success of the QWERTY keyboard and the VHS tape format were dependent on their particular historical paths and was not a simple reflection of product optimality or superiority.
Robin Cowan has also published several compelling examples of path dependency (1990, 1996). In his 1990 essay, he put forward a detailed overview of the development of nuclear power reactors in which he argued that light water reactors emerged as the dominant technology despite the fact that it was "not the best technology, either economically or technically." He concluded that the "technology which first makes large advances along its learning curve will emerge dominant."
The key explanatory variable in these types of stories is the presence of network effects: "Consumers often place higher value on a product if other consumers also use it. When this occurs, the product is said to display network effects or network externalities" (Besanko et al. 2004). In this view, there are two types of networks, actual and virtual. Actual networks physically link consumers or users together, as with telephone and e-mail systems. The more users there are in these actual networks, the more value consumers or members derive from participating in these systems, and this value is a reflection of the actual network effect. In virtual networks, users are hOt physically linked so "the network effect arises from the use of complementary goods. Computer operating systems, video gaming ... and DVD players" are leading examples of networks in which member utility or value increases as the number of users of these systems rises.
In a more general discussion of this process, Michael Katz and Carl Shapiro write that "[t]here are many products for which the utility that a user derives from consumption of the good increases with the number of other agents consuming the good.... The utility that a given user derives from a good depends upon the number of other users who are in the same network" (1985). This suggests that sometimes consumer utility depends not on how good a product is but on how many people use the product: popularity may prove more important than usefulness or effectiveness.
A contemporary example of technological lock-in can be found in the new productivity upgrade from Microsoft Office 2003. In order to get the full functionality from Office 2003, a user must also adopt Microsoft server technology and architecture. A user who has made this adaptation is locked in (and competitors such as Sun and Oracle locked out). As a result, the idea of technological lock-in is well known within the information technology industry.
There is, however, another dimension to lock-in that does not depend on technological networks, yet it is not as well documented in empirical work. We refer to this category as behavioral lock-in. Once a product has become established as an industry standard, and once consumers or users have invested rime or money in learning a particular system or becoming comfortable with a traditional practice, they will be less likely to try a rival process, even if over rime it proves superior. This is not necessarily a new concept. In his 1985 essay, David asserted that path dependencies may arise "in the presence of strong technical interrelatedness, scale economies, and irreversibilities due to learning and habituation." However, nearly all essays on path dependency have focused on technological lock-in; the term behavioral lock-in is meant to stress and develop David's concept of "irreversibility due to learning and habituation." Behavioral lock-in occurs when the behavior of the agent (consumer or producer) is "stuck" in some sort of...