Kathleen Thelen, How institutions evolve. Insights from comparative historical analysis” in Mahoney and Reuschemeyer (eds.) Comparative Historical Analysis in the Social Sciences (2003), Chapter 6.

Summary: Thelen argues, in contrast to the punctuated equilibrium models that are advocated in path dependency theory, scholarship has to acknowledge how institutions are renegotiated periodically without drastic change (213), also called “constant causes”.  Thelen proposes framing issues using two concepts – institutional layering (partial renegotiation of some elements of a given set of institutions while leaving others in place) and conversion (where existing institutions are redirected to new purposes, driving changes in the role that they perform or the function they serve) (226).
Method: Thelen outlines a categorical framework for institutional evolution arguments through three approaches to institutional creation and change: functionalist, power distributionist, and cultural, including brief country-based case examinations as examples.
Important Insight: Aptly, Thelen recognises the role of “losers” in the institutional evolution process as actors who do not necessary disappear, rather they adapt while perhaps not “embracing and reproducing” institutions (231). In doing so, Thelen recognises that institutional evolution and activity is a “repetitive game” in the rational choice language.
Critique: At the outset of the article Thelen states that the critical junctures or exogenous shock explanations of institutional change are more prevalent than the “constant causes” explanations, however, she does not discuss the use of exogenous shocks by the elites to mask change through institutional layering and conversion.
*key point: In order to appreciate fully the causal weight of the past, we need to complement analyses of path dependence and punctuated equilibrium models with conceptual tools that can capture the logic of institutional evolution and change. Specifically, we need to take account of the way in which crises or turning points can be generated endogenously, including through the working out of the logic of previous switch points. We also need to consider the way in which new problems are conceived and solutions sought are themselves products of the past rather than historical accidents (Haydu 1998)
*This amounts to a call for introducing somewhat more structure at the ‘front end’ of the analysis of institutional development than most path dependence arguments do – by attending to the way in which historically evolved structures limit the options of political actors even at critical choice points. It also calls for injecting somewhat more agency and strategy at the ‘back end’ of such arguments – by emphasizing the way in which institutions operate not just as constraints but as strategic resources for actors as they respond to changes in the political and economic contexts that present new opportunities or throw up new challenges.
*The issue of institutional change is an enormously complicated one, and my aim in this essay is relatively modest. I introduce two ways of conceptualizing the problem of institutional evolution that overcome the current zero-sum view of institutional innovation versus institutional reproduction that has been characteristic of much of the literature in this area to date.
*Key Contribution: Arguments about institutional change through layering and through conversion incorporate elements of increasing returns arguments from the path dependence literature, but they embed these elements in an analysis of ongoing political contestation over institutional outcomes. In doing so, they highlight the processes through which institutional arrangements are renegotiated periodically in ways that alter their forms and functions.
*ongoing theoretical work centring on the concept of path dependence by Mahoney, Pierson, and others has lent greater precision to pervious formulations based on the dual notions of ‘critical junctures’ and ‘historical trajectories’ (Mahoney 2000; Pierson 2000)… But path dependence as currently conceptualized tends to encourage a rather strict separation of the issue of institutional innovation and institutional reproduction. Much of the work that invokes this concept is premised on a punctuated equilibrium model that emphasizes moments of ‘openness’ and rapid innovation followed by long periods of institutional stasis or ‘lock in’ (e.g. Krasner, 1988). The implication is that institutions, once created, either persist or break down in the face of some kind of exogenous shock… But what about institutional changes that fall short of breakdown (e.g. persistent communist era institutions in post-communist states), or the large cumulative effects of ongoing but often subtle changes that persist over long stretches of time (e.g. British House of Lords)?
*the contemporary literature on institutions does not yet offer very good conceptual or theoretical tools for addressing these phenomena… Arguments about dynamic processes (e.g. Pierson 2000; think increasing returns arguments) tell only part of the story; they are better at articulating the mechanisms of reproduction behind particular institutions than they are at capturing the logic of institutional evolution and change. Moreover, in some cases, explaining institutional persistence itself may require us to look beyond arguments about increasing returns (specifically re: elements of institutional transformation designed to bring institutions in line with changing political, social, and economic conditions).
*prominent approaches that view institutions as brought about and sustained by particular, specified causal factors (with changes in the particular foundation of an institution resulting in institutional change):
**utilitarian-functionalist: functional requirements as the causal factor
**power-distributional: an underlying balance of societal or political power as the causal factor
**cultural-sociological: a shared understanding of what organizational forms are considered legitimate or appropriate.
*path dependence perspectives draw on the same causal arguments as the first three perspectives, but where they differ is that they do not necessarily look for ‘constant causes’ through time and space. As Mahoney points out (2000, p 515) path dependence perspectives posit that factors responsible for the genesis of an institution may well be different from those that sustain it over time… Institutional change is often seen to be a function of exogenous shocks that disrupt previously stable arrangements and open the door again for institutional innovation.
*two potentially fruitful ways forward in this debate (there are many others), which have the benefit of avoiding facile functionalist approaches that read the origins of institutions off the functions they currently perform, and are genuinely historical, looking at social processes as they unfold over time and in relation to other processes:
**layering involves the partial renegotiation of some elements of a given set of institutions which leaving others in place (e.g. US congress, where new coalitions which lack the support necessary to demolish an institution may simply erect a new one, see Schickler 2001)
**conversion involves the redirection of existing institutions to new purposes, driving changes in the role they perform and/or the functions they serve (e.g. US poverty programme used by Johnson to deal with racial tensions, see Weir 1992).
*tasks for future research:
**we must ask ourselves, even in cases of increasing return and lock-in, whether it crowds out layering or conversion.
**we must distinguish the types of empirical phenomena that are associated with the different modes of change (i.e. layering and conversion).
**in general, it may be unwise to draw to sharp distinctions in history b/w ‘settled’ and ‘unsettled’ times. Instead, to understand how institutions evolve, it may be more fruitful to aim for a more fine-grained analysis that seeks to identify what aspects of a specific institutional configuration are (or are not) renegotiable and under what conditions.