Table of Contents
Path Dependency as a theory postulates that initial decisions, conditions, and managerial directions are likely to affect subsequent decisions or concurrences. The theory argues that history has a central place in influencing the decision-making in an organization. Vergne and Durant (2010, p. 736) further explain that this theoretical assumption aims to connect the past and future in an abstract manner. Moreover, in its most minimal form, path dependency is manifested through the persistence of a decision, or its apparent durability over a historical epoch (Margolis and Liebowitz 2011). However, Sydow, Schreyögg, and Koch (2009) ascertain that there is no widely encompassing yet particular enough definition of path dependency in the way that would serve to make it a theory per se. Thus, according to these scholars, the concept of path dependency is more figurative than theoretical (Sydow, Schreyögg, and Koch 2009). On the other hand, Liebowitz and Margolis (2001) have even questioned whether path dependency exists at all since drawing upon these researchers, it is impossible to stay locked into an inferior choice when one can choose a significantly superior one. Nevertheless, an analysis of path dependency reveals that it has four major elements. These elements have a major impact on the social science research leading to imprinting, potentially incorrect results, and lack of homogeneity in the some of the results.
The Core Elements of Path Dependency
The first element of path dependency emphasizes that there is a certain level of non-predictability at the start of any process that ends up being path-dependent. This circumstance might come out as a contradiction of the theory because the premises of path dependency deal with fixed, or quasi-fixed, outcomes in the way that the people and organizations make a decision. However, non-predictability implies that during the first instances any process that later becomes path-dependent, several choices are usually in the market, with a possibility of the social or economic market adopting any of the choices. Nevertheless, this issue does not imply that the choices have to come into existence at the same time. One of the choices may suffice before the others do, but at one point, consumers must have a definite choice on the trend or the product they want to consume. In spite of this, however, consumers choose one of the trends or products over another. The choice of the product by the market may not be dictated by its utility over another, but may be dictated by non-market forces. The choice occurs at a critical juncture and could be a decision or an accident. The case of a critical juncture that is an informed decision is apparent in the example of the person, who in spite of being aware that the locality one is buying a house might be in the vicinity of a future sewer treatment plant. Nonetheless, a consumer buys it in order to be closer to friends and family. Taking into account the same scenario, an accidental critical juncture is that one might buy a house in an area unaware that several years later, the city government will decide to build a sewage treatment plan in the vicinity of such a house, and thus, end up owning a house with little market value not due to the person’s informed choice but an accident.
The next characteristic concerns the quality of non-ergodicity. A system is ergodic if, in its path of development, it is clearly independent of the conditions that it has obtained at initial stages. Non-ergodicity is the antonym to ergodicity and this concept refers to the breakdown of ergodicity in any system. This circumstance leads to the process becoming path-dependent. In this case, choices that organizations make long after the transitory conditions impact the nature of the future conditions as “history matters” (Schreyögg, Sydow and Holtmann 2011, p. 82). Such a situation implies that in some cases, it might be out of order to use ensemble data to make inferences about individuals while conducting social science research. For this reason, opinion polls might be utterly wrong about, for instance, an election result as the data and the margin of error are collected with the assumption that the human sample used (the ensemble) are ergodic while the opposite might be true. Thus, in social science, if scientists conduct a generalization towards an ergodic sample, the generalization is likely to be correct. As such, path dependency dictates that if the generalization from a particular data set towards the general population is non-ergodic, it is likely to be incorrect. Moreover, non-ergodicity implies that from the multiple equilibriums, history will select he most appropriate outcome from the several outcomes that are possible (Sydow, Schreyögg, and Koch 2009, p. 690). The outcome itself is related to other past developments. Based on the nature of the outcome, one cannot consider it apart from the past.
The third aspect of path dependency is inflexibility. In most cases where path dependency is apparent, inflexibility is present in the outcome. This factor is due to self-reinforcing mechanisms, namely, it is likely to follow a certain route regardless of the nature of means an actor in the system might take in order to avoid a particular outcome.
The self-reinforcing mechanisms develop due to four major effects Sydow, Schreyögg and Koch (2009) have identified. The first is the coordination effect since in a particular system, the interaction among the various actors will be efficient as long as they apply a single system: for example, using right-hand as opposed to left-hand vehicles. The next one concerns the effects that come as a result of complementarity. In this case, the advantages can be linked to rapidly combining the interrelated activities as a form of synergy. The third mechanism concerns learning effects. As one performs a process more frequently, he or she becomes more skillful in doing the task, and thus, efficiency is gained. The fourth effect is the adoptive expectation effect that concerns a person’s change to fit in the social conditions. This circumstance leads to the propagation of certain products as more people get a particular product, the more useful it becomes. An example is a telephone whereas its utility increases since more people in one’s social circle get one, including family and friends. These self-reinforcing mechanisms can lead to a lock-in. With time, they become systematic forces that cannot be changed by individual actors. Therefore, an individual actor is entrapped in the dynamics of the system from where they are unable to extract themselves from (Sydow, Schreyögg, and Koch 2009, p. 694). This inflexibility could thus be dictated by the predominant social influences that might leave some space for limited variation in the outcome (Sydow, Schreyögg, and Koch 2009, p. 695). However, such variation will only occur in the case where there are knowledgeable agents who interpret the “path” in the organization differently and will only occur in the performative side in spite of the fixed overarching organizational pattern.
Moreover, the lock-in has a major implication on efficiency. This situation results in the fourth edifice of the theory, namely inefficiency. The lock-in rarely leads the most efficient choice. On the contrary, the market, either economic or social, is locked into the inferior solution. Nevertheless, at the initial stages of the development of a social event, path-dependent processes might lead to path-positive feedback. This circumstance is highly influential in the social adoption of processes that may later become locked-in as the positive feedback leads to a diminishing possibility of choosing an alternative later in the process. Greener (2004, p. 617) gives an example of how the above procedure occurred in the late 19th century: Germany quickly overtook the United Kingdom in industrial development in spite of the UK having several years head start. While Germany was relatively new to industrialization and could, thus, bypass some of the older technologies, the UK was locked-in technologies that had been useful a few decades before but had become obsolete and were also too expensive to replace. For this reason, replacement of the technologies would have led to wholesale scrapping of the industrial machinery as it was applicable at that time though this approach did not appear viable, at least in the short term.
Sydow, Schreyögg, and Koch (2009, p. 695) have noted that lock-in does not have to result in an inefficient outcome. However, this assumption ignores the fact that inefficiency in path dependency does not have to refer to the present time. Moreover, with the human inability to correctly predict future outcomes, rigidity at present has the effect of potentially impacting future outcomes negatively. Specifically, in future, newer challenges will crop up that the organization will need to confront in a manner that is disparate from what the set path has anticipated. Apart from that, even if the same challenges remain, better, more dynamic solutions might be available that the organization will not be able to use due to the lock-in. This issue is by definition inefficiency as it binds the organization to the historical way of solving issues which has been overtaken by better methods. Additionally, thhis inefficiency impacts an organization’s ability in future to create new ideas. This shift in rationality or dysfunctional flip has an effect of slowing down the entire system or organization. Since inefficiency is only measurable in comparison to another system or organization, this introduces the case for comparative analysis in social sciences. However, in light of such comparison, there has to be a base of reference. Therefore, for social scientists, the question is how to choose the referential base without eliciting calls of preference for a certain outcome, especially in both qualitative and quantitative research.
Impact of Path Dependency on Social Scientific Research
In its most basic of forms, path dependency implies that history will determine how human beings make and implement present decisions. By its nature, social research will concern the study of human life, interaction, and organization. The first impact concerns the fields of sociology and organizational theory. Social scientists have developed the concept of “imprinting” that seeks to explain how the initial condones in the environment leave an “imprint”, namely, a mark that it persistent in communities. Thus, social scientists, while studying the behaviors of these communities, later on, have to consider the impact the environmental conditions might have had on such a behavior as it persists long after the environmental conditions have been established.
Secondly, the theory of path dependency means that scholars in the field have to accept that some studies in social science can have incorrect results. This would be despite that the social scientists have followed the scientific method to the core. The incorrect result is because of their non-ergodic nature, especially in the studies whose main goal of studying the human behavior using samples and extrapolating them to the entire population and then back to the individual. For instance, if scientists conduct opinion polls, they gather data from few individuals and try to infer the characteristic of the particular populace. The individuals are picked based on the particular criteria by making sure all the groups are represented. They then calculate the margin of error that again has a basis that the ensemble they have created is an ergodic while, in the real sense, it is non-ergodic. This is the same issue that other social scientists face every day when trying to infer a statement that is general from an experiment that has been very particular. The ergodicity, or otherwise, the generalization might determine the correctness of the whole result.
Thirdly, as Lambert (2014, p. 131) notes in his study, social scientists should note that there is a lack of homogeneity in the success or failure of various players in the market. While the playing field, probably, is equal, the past circumstances of the various players are not, and as such, their managerial decision will be different over any particular time.
One of the most cited examples in Path Dependency theory is that of the QWERTY versus the Drovak keyboards (Vergne 2013, p. 1193). The QWERTY keyboard has been dominant for more than 100 years and is likely to remain so for a long time in spite of the development of new, more technically efficient keyboards throughout the years, including the Drovak keyboard (David 1986). This issue has mainly been advanced by the network effect because the more the people who had typewriters with the QWERTY keyboard, the more the companies could do a mass production of the typewriters. Thus, there has been a technological lock-in of more than 100 years even though there is no evidence that QWERTY keyboard is better than its alternatives.
Path Dependency as a theory stresses on the importance of past actions and decisions on the current ones. The theory has four major elements: non-predictability at the initial stages and the later stages; non-ergodicity, non-flexibility, which in turn, leads to inefficiency. This factor is apparent in the case study of the QWERTY keyboard. Path Dependency has a major impact on social science research due to these characteristics. First, it has led to the development and the acknowledgment of the concept of imprinting in social science research. Secondly, this hypothesis has led to the acknowledgment that some of the results of social science research could be wrong if they involve the extrapolation of the general from the specific. Lastly, path dependency has led to understanding that the lack of homogeneity among similar phenomena in social science research is preventing as their historical backgrounds are.