“At every step along the way there [are choices]—political and economic—that provide…real alternatives. Path dependency is a way to narrow conceptually the choices set and link decision-making through time. It is not a story of inevitability in which the past predicts the future.” -Douglass North
Methodology is the means in which to study and determine how and why a phenomenon occurred. Social scientists use rigorous deductive logic to develop both intuitive and counterintuitive hypotheses about the dynamics of causal mechanisms. There are three predominate methods used in the social sciences today, they include: using statistical analysis as a purely quantitative approach; two, using case studies in a comparative perspective; and third, and finally, creating formal modeling (George and Bennett, 2005). Inductive logic is used for theory building. Etymologically these are all the same, you look at what are the implication of actions, but the focus, however, is still on how you go about to get them.
For the majority of the empirical methods class we have focused the controversy started by King, Keohane, and Verba’s (KKV) book Designing social inquiry (1994). In which the authors try to promote social inquiry used in statistical analysis and apply it to case study research. By doing so, KKV set to encourage hypothesis testing, increase large N analysis, explore causal mechanisms and promote conceptual validity as a means to avoid case selection biases and non-rigorous qualitative research. Furthermore, they supported the decreasing degrees of freedom as a way to promote tighter or more rigid analysis within case studies approach. Although this application of quantitative approaches into qualitative studies provided a structure for debate, it also developed into a large body of criticisms by many political scientist, international relations specialist, sociologist, and anthropologists, who view KKV as inaccurate and an over simplification of their work.
As a result, the class has focused on the application of various methods used as a criticism or an alternative to KKV’s assumptions. For example many social scientist have criticized KKV for the way they have conceptualized descriptive, causal inference, and causal mechanisms, as well as concept formation and the methods of measurement within comparative methods and case study analysis. Additionally, the class looked at ways to select cases, including developing typologies or narrative cases and using preventative measures for selection biases such as not claming “too much”. Furthermore, we have learned methods “within cases” such as process tracing, congruence testing, differences between differences, and examined how to develop a counterfactual analysis. Temporality, critical junctures, and path dependency are just other empirical methods to understand why phenomena occurred, but it uses time as its key unknown quantity or condition.
Path dependency, which will be the focus of this paper, provides another way to look at what is happening within the black box and sheds light why possible outcomes occurs, even if they are not always the most efficient. It describes why history matters and provides information about the possible sequencing of events. This method provides in-depth rational as to whywhy things occur and not whether their historical background impact, but how it does. Its main goal is to inform what is going on within and between independent variables in order to better understand empirical tests of causal mechanisms. independent variables matter. Developed further in the paper, the use of path dependency in case study analysis is not just concerned with inputs and outcomes, which are so provided in a typical quantitative study—defined by identify a dependent variable and regressing them on the most valuable independent variables, watching out for heteroskasticity, multi-colinearity and/or omitted variables. But rather, this method seeks to identify
The reason why I have selected Temporality, Critical Junctures, and Path Dependency for the topic of my critique essay is that I believe that it is ultimately impossible to completely isolate independent variables. Whether an empirical test has carefully and meticulously selected its independent and dependent variables, there will always be a place for spurious and or omitted variables arguments because of interaction between the independent variable in society, which are so difficult to separate. Social systems at the micro-level are constantly bleeding and blending within their political, economic and cultural systems. No one social system is neither ubiquitous, nor is one solely independent, rather we are all interdependent. Yet, an essence of time can provide a social scientist with a constant in which to provide comparisons.
Rueschemeyer and Stephens (1997) argue that comparative research can benefit from historical sequences, when it takes into account the varied historical consequences of a particular event. Furthermore, a study which uses this comparative approach can help with macro-sociological studies by eliminating problems like the small n, lack of independent cases and provide a more solid explanation of their association for further causal exploration. They suggest that the uses of historical accounts are important because phenomena are shaped by environment. Oftentimes the sequences of events may determine its causation. The authors suggest that its better to know the history than just a few data points without the context.
Yet, the authors know that history can also be subjective. Especially when social scientist look into only relevant evidence and operationalized variables measured at the appropriate time. Although it is argued that this helps to avoid multicollinaritity among independent variables, even better than with pool data cross-sectional with time series. It may also create questions of spuriously correlated results. Finally, Rueschemeyer and Stephens suggest that there can be different theoretical explanations for a given correlation when analyzing vary complex patterns of historical propositions. There is not just a single correlation but complex patterns of segments in different national contexts. They conclude by suggesting that “Data analysis simple tweaks theory and one must go back to the qualitative research to also verify if it’s correct…. This is the difference in deterministic vs. probabilistic theory, when selecting a single case…. History was not predetermined…. Cases not a single dot on a scatter gram, but always a deviate case that may be contingent on an open variation” (Rueschemeyer and Stephens 1997: 69).
Pierson (2000) furthers their analysis of why history maters and suggests that its not just timing but also sequencing that is important in understanding a phenomenon. Pierson’s almost obsession with path dependency rails other political scientists over-usage of rational choice theory for the explanation of social and political phenomena. Pierson cites Charles Tilly (1984) statement “when things happen in a sequences, it affects how they happen.” An important observation by Pierson and others is that sometimes the most logical outcomes (with rational decisions made) are not a product of the process.
In his first article, Pierson describes sequencing, path dependent processes, non-reinforcing events, historical conjunctures and temporal effects. His second article describes economic and political perspectives of why increasing returns and path dependency are so important. Finally, he describes historical institutionalism and provides an analysis of why history matters for policy studies. This paper will summarize his arguments by using Arthur (1989) Mahoney (2000) and David (1985) for more information. In addition, it will present Kay (2005) counter argument for the use of path dependency in policy studies and finally, I will present my perspective to this debate.
Path dependency is at work when there are self-reinforcement or positive feedback loops. Once something occurs its goes down a particular road and does not take another route because it becomes intertwined with the first decision. Furthermore an event may be iterative, but an analyst can always return to the critical juncture—or the starting point of when a decision was made. Pierson then argues that when a particular event within a sequence takes place, it also makes a difference. As suggested in class with that example of the WMDs in the Iraq war, “small” events early on my have “larger” impacts, while large events later may be less consequential (Bossinger). Pierson cites Mahoney, “these feature of comparatively ‘open’ processes at the outset, followed by a more constrained choice-set once reinforcement sets in, are precisely the features that make this type of process distinctive and underpin the claim that temporal sequences is crucial” (Pierson 2000:75).
Pierson further suggests that there are various sources of path dependency. He builds this argument from Brian Arthur’s study of technology and the evaluation of increasing returns. First, there are large set-up fixed costs; next possible learning effects; later to have coordination effects; and finally, the adaptive expectations of a new invention/ product that goes into the market place. The example used is from David (1985) with the story of the adoption of the QWERTY keyboard. The background describes why the DSK version of the keyboard that Dvorak and Dealey patented in 1932 was never massively adopted within society even though their system was much more efficient form of typing. The story identifies critical conjectures taken within the US and Britain and why the socialization of the more efficient keyboard was not adopted. Pierson later identifies other technologies have had the same fate as the keyboard such as “the battles between Betamax and VHS video recorders and between DOS-based and Macintosh computers, earlier automobile designs, and competing standards for electric current” (Pierson 2000: 254). Therefore this argument of path dependency as key for the increasing returns of technology within micro-economic theory can also be transcribed into political world.
Mahoney furthers that self-reinforcing sequences once inserted into institutions, “as enduring entities that cannot be changed instantaneously or easily” modified create a locked-in effect (Mahoney 2000:512). Once a critical juncture has created a lock-in, a researcher can explore other outcomes for example the “what if” of counterfactual analysis. Mahoney provides a typology of path dependent explanation for institutional reproduction. He suggests that the mechanism of reproduction, potential characteristics of institutions, and the mechanism of change are different on the various explanations of utilitarian, functional, power or legitimation of a particular outcome. This identifies the type and how “tied” the path of locked-in outcome might have. This typology provides a way to think about when path dependent outcomes interact with an institution. This is easily seen in daily life for a public administrator, who knows that once rules have been in place, they fill them within their jobs. If some individual wants to change how things are run within the institutionalized system, it becomes vary difficult for them.
Mahoney later defines that two independent sequences of events may have
“no effect.” While a conjuncture may have no enduring consequences, other conjunctures may have a lasting effect and thus enduring consequences (Mahoney 2000:529). These critical junctures are what analysts should focus their attention. That moment within the sequences is important because those junctures create different outcomes. Temporal effects may not have had lock-in where the conjuncture occurred. Therefore large events studied should be large enough for greater change to happen within a system.
With that, Pierson furthers his argument of why path dependency is important to the political world. He highlights that timing is important for when public policies are implemented and introduced into the system. Political resources are often limited with many constraining factors. For example, time constraints of politicians within the election cycle depend on when they can maximize their opportunities. Another application is John Kingdom's often cited quote “the window of opportunity” for when a policy change enters into the political environment. Furthermore, Pierson argues that the complexity of power authority and power asymmetries also complicate path dependency with policy studies. He suggests the there is a collective action problem within politics, which is much like a common-pool resource in economics where everyone benefits from the public good. Yet, the struggle of politics is much different than economics, its much more complex, lumpy and nuanced.
Adrian Kay critiques the use of path dependency in policy studies because he suggests, although it’s fashionable to label history mater, without clear and convincing accounts of the decision-making process, there can be no understanding of a phenomenon (Kay 2005). He suggests that it’s hard to operationalize temporal dynamics that operate in a policy world. Contrary, he explains stability and not change should be an analyst focus. Emphasizing that point, he reiterates that the normal implication of social life are confused and often left unexplored. Whereas the ideas of path dependency are good, he adds, “We want to move to a world of moving pictures not static snapshots.” (Kay 2005) Furthermore, he inserts that the functional moves of how thing happen and when they occur affecting their trajectory is undeniable. Kay finally argues that path dependences isn’t useful for policy studies because it lacks convincing accounts of decisions making over time, it has bounded rationality and it lacks a normative focus.
These arguments ultimately fit into the broad institutionalist framework developed by North set at three levels: the macro conditional level; the collective degree or policy decision making level; and finally, the operational level of individual decisions people make on the ground. Kay suggests that the theories of path dependency hint at the macro-environment but when he tries to ground them into the mezzo level, their meaning is somewhat lost. This ties into the lock-in of an outcome once an institution has adopted a particular policy, and later, if it must be change creates complex problems for possible remedies.
Finally, Kay highlights Pierson suggestion that most people focus on macro theoretical level, but what is key is to look at the policy level. Once an outcome becomes taken over by the public administration/ institution its outcomes are locked-in. The institution takes over and it’s collectively accepted. The adapted rules or constrains of a particular outcome are then governed by the behavior of the institution. Therefore, Kay suggests, path dependency might help to support theory building but it’s hard to test empirical policy work. This is relevant since there is little empirical evidence to-date, which uses this method. Finally, Kay emphasizes the application of a theorized concept of a policy development might be sui generis, that it is un-functional to anything but its self.
Kay misses the policy process as a particular way to evaluate path dependency. This messylumpy is hard to judge. arrange of interests that Pierson highlights as being Pierson suggests politics does not have the simple measurement of “price” to evaluate political interests. I would agree that once a policy has been institutionalized (and become a self-reinforcing process), by becoming law, funded by congress, or a regulation set within the bureaucracy, it becomes very difficult to change. But the political realm of making decisions is where political scientist focuses their attention. This messy environment is exactly what the studies of path dependency do to help study why a particular decision was made and why another maybe more efficient decision was not.
But my critique of the use of path dependency for a method of political analysis is also mixed and somewhat nuanced. I would argue that there is subjective bias in case selection of critical junctures of a particular phenomenon, which an analyst selects for various reasons to study. Yet, the last effects of research may create greater understating of why and how things occur, but they may not be able to be replicated into the future. An example, which came to mind, is that of Graham Allison’s study of Kenney’s decision making progress for the Bay of Pigs. No doubt, did this single event create a critical juncture in the Cold War and could have aided or explained the prediction with the arms race the US had with Russia in the 1970s and 1980s. But as North suggest in the introduction quote, “no one can predict the future” (Pierson 2000).
Yet, the subjective bias of case selection is the nature of social sciences. When you select a particular case, and which ones to comparative it to, is ultimately the trick. What KKV suggest is that you can create larger N for comparative study or you can apply various cases into a typology for a more robust analysis. These techniques needed to be done within the political realm of policy studies to have a better understand of how path dependency works.
Path dependency, as an empirical method, does describe how to select points in an analysis link together and it does describe the independent variables and how they interact within a model. The method is successful to directly attacking rational choice as the “only” plausible outcome of a phenomenon within economics and political science. There, of course are more nuances to be understood and more empirical studies need to be taken into consideration before one can draw final conclusions.
Arthur, B. (1989). Competing Technologies, Increasing Returns, and Lock-in By Historical Events. Economic Journal, 99(394), 116-131.
David, P. (1985). Clio and the Economics of QWERTY. American Economic Review, 75(2), 332-337.
George, A. and Bennett, A. (2005). Case studies and theory development in the social sciences. Cambridge, MA: MIT press.
Kay, A. (2005). A Critique of the Use of Path Dependency in Policy Studies. Public Administration, 83(3), 553-571.
King, G., Keohane, R., and Verba, S. (1994). Designing social inquiry. Princeton, NJ: Princeton University Press.
Mahoney, J. (2000). Path Dependence in Historical Sociology. Theory and Society, 29(4), 507-548.
North, Douglass. (1990) Institutions, Institutional Change and Economic Performance, Cambridge: Cambridge University Press.
Pierson, P. (2000). Not Just What, but When: Timing and Sequence in Political Processes. Studies in American Political Development, 14, 72-92.
Pierson, P. (2000). Increasing Returns, Path Dependency, and the Study of Politics. American Political Science Review, 94(2), 251-267.
Rueschemeyer, D. and Stephens, J. (1997). Comparing Historical Sequences: A Powerful Tool for Causal Analysis. Comparative Social Research, 16, 55-72.