Kathleen Thelen. How Evolve: The Political Economy of Skills in ,Britain, the and Japan.New York: Cambridge University Press, 2004. 352 pp. Index. $75.00 (cloth), ISBN 0-5218-3768-5.

Reviewed by Gary Herrigel, Department of , University of Chicago. Published by H-German (January,2006)

Untitled

This is a superb work of comparative labor into the governance of the (extended) historical political economy.Itmakes a sound and vocational system. In Japan, it meant replacing the enlightening empirical contribution to our (neo)artisanal okayata trainers with an in-house understanding of the emergence of four quite system of training structured by a seniority system distinct national systems of vocational training. and permanent employment and in which company And it makes a provocative theoretical argument unions assumed the status of stakeholders in about the character and process of institutional enterprise. While none of the developments in change overtime. This reviewwill deal with each either case occurred easily or all at once (Thelen of these dimensions of the book in turn. tells very detailed stories about howeach system The empirical arguments about the four developed), in both cases, the crucial detail is that national cases are synthetic and rely,for the most the interests of organized labor and employers part, on already existing secondary sources in each tended to run in the same direction. of the countries. Nonetheless, the comparative In areas where the line between artisan and causal arguments about the factors that gav e rise to industry was blurred, however, unions focused on each of the systems are remarkably subtle, original, the control of skill and apprenticeship as a way to and very illuminating. The key actors in all four secure control overthe labor market. As a result, stories are large metalworking firms--big training became a central focal point of class demanders of skilled labor at the beginning of the struggle with employers. In Britain, conflict twentieth century--and trade unions. The crucial between craft unions and employers overcontrol of structuring difference shaping the terrain of the production process profoundly shaped the possibility for action in all cases was the ev olution of training practices. Advances in relationship between the traditional artisanate and training often occurred under conditions of market the state. shortage or governmental enthusiasm, but then In Germanyand Japan, state policycreated were rolled back when those conditions changed. (or protected) a clear line between traditional Over time, this produced a highly diluted and artisans and skilled industrial workers, while in inconsistent training system that became weaker as Britain and the United States state policybroke the twentieth century progressed. In the United down that distinction. Where the line between the States, competition between unions and employers artisanate and industrial workers was maintained, for control overthe workplace was, if possible, trade unions and employers did not struggle over ev enmore intense than it was in Britain. But issues of training (because it was the province of a unions in the United States were weaker than their different social class), and instead both focused on British counterparts and employers aggressively optimizing, supplementing or extending the sought to uproot them from the shop floor by artisanal system. In Germany, doing this meant replacing skilled labor with automated machinery extending to industry the in-plant training system and unskilled labor.Overthe long term, as in developed within the artisanate, allowing industrial Britain, training deteriorated and neverbecame employers to certify skilled labor along with the institutionalized in anybroadly effective or artisanate and, ultimately,incorporating organized systematic way in the national political economy. H-Net Reviews

Crucially,Thelen shows howthese different character of institutional transformation overtime. outcomes in the four cases affected the Here the argument is very ambitious, intriguing, attractiveness to youth of investing in training. In butalso, ultimately,limited in a variety of ways. Germanyand Japan, where the training systems ForThelen, the issue behind the issue in her had great legitimacyand institutional robustness, portrait of vocational training systems is the becoming an apprentice and engaging in the constant character of institutional change, evenin training system was rewarding to both the periods of apparent or alleged stability.She is apprentice and the employer.But in the United dissatisfied with the traditional argumentation in States and Britain, where there was no systematic political science/political economy that separates apprenticeship system with reliable quality,youth the problem of institutional change from that of ev entually showed very little interest in investing institutional reproduction. In that literature, change in formal training. Due to differences in tends to be understood as something discontinuous, institutional structure, in other words, the content provokedbyanexogenous shock that either wipes of skill, its general availability,and its social status out institutions, or so destabilizes them that it were constructed differently in each of the political creates the possibility for previously unforeseen economies. agencyand innovation. Reproduction, on the other This is a very elegant institutionalist hand, is what happens in those periods when there argument. Indeed, it is perhaps too elegant in that it is not change. It is characterized by stable positive makes the "systems" that it describes more feedback mechanisms that sustain the integrity of inclusive and coherent in each of the societies than institutional systems overtime (path dependency). theyactually were/are. In some cases, Thelen Using the German case, Thelen shows that recognizes that alternative practices existed (at this hard (conceptual and temporal) dichotomy least in the past), but she dismisses them as between change and reproduction really cannot be exceptional or niche forms that were either sustained. Seen overthe course of nearly a century, absorbed or supplanted by the system she describes the German case exhibits both remarkable (pp. 49-50; 186). She tells a teleological story in continuity and significant change. From its which the emergence of a single system, and the inception in the 1890s, the German vocational actors associated with it in the present, erase the training system has always been a "collectively social and historical memory of alternativestoit, managed system for monitoring howfirms train rather than an open ended narrative inwhich their workers" (pp. 296)--this despite the multiple practices reproduce themselves in greater experience of several significant social, economic, and lesser accord with evolving formal rules and and political crises and breaks in the society.But institutional design. This orientation comes back to the system has been governed in several different haunt her in her discussion of the current crisis in ways overthat same period. Moreover, the the contemporary German vocational training intentions behind the construction of the initial system--she essentially sees crisis, but is at a loss arrangements were in manyways diametrically as to conceive ofpossible alternative arrangements opposed to those that sustain very similar in vocational training (pp. 269-277). By arrangements today.Inproviding the artisanate eliminating the noisy alternative practices from her with monopoly authority to certify apprenticed narrative,she has nowhere to look for experiments training, the German state at the time hoped to about alternativestothe system in crisis. bolster that petty-bourgeois class and undermine That said, Thelen’sbook will stand as an the momentum of the growing labor movement. It indispensable resource for anyone wishing to wasananti-trade union, anti-social democracy understand howvocational training in each of the measure undertaken by a conservative regime four national systems emerged, howitworks and struggling to contain growing demands for howskill is defined, organized, monitored, and representation. Yet by the mid 1990s, that same understood in dramatically different ways in the system of collective management and monitoring different national institutional contexts. of skills in Germanyfully incorporates the But Thelen is not merely interested in participation of labor unions and was widely addressing an international community interested viewed as a core in the labor-friendly in vocational training. She also makes a very subtle German production system. Thus, despite and interesting theoretical argument about the remarkable continuity overacentury,there was

-2- H-Net Reviews

profound change in the system. Institutional framework that takes account of the feedback reproduction and institutional change are deeply effects that have defined the conditions with which intertwined in this history. specific policyand institutional choices are being Howtoconceptualize this type of gradual made. Theyhighlight the way that policies initiated change within continuity? Thelen takes two at one point affect which actors are around to fight conceptual steps to try to do this. First, she brings the next battle, howtheydefine their interests, and agency-based explanations together with structural howand with whom theyare likely to ally ones. She praises, for example, the work of Peter themselves subsequently"(pp. 288-89). Swenson who emphasizes the crucial role of Both views, in other words, are historical: capitalist agents in constructing institutions for theythink that it is indispensable to understand domesticating market dynamics in contemporary present action in the context of past actions. And capitalism. For Swenson, institutions represent the both views understand the historical process as solutions to specific kinds of environmental characterized by feedback effects (past actions problems confronting actors. Previous solutions shape future possibilities). In her rendering, the create the conditions for choice in a subsequent structuralists have the virtue of defining the round. Thelen, however, also acknowledges very historical terrain very broadly--beyond the purview favorably the criticism of Swenson’swork by of individual agents. Larger context constrains the structuralists such as Pierson and Hacker and agent by limiting the choices available. It also, Stephens and Stephens who emphasize that agents however, can change in ways independent of a do not construct institutional solutions out of particular institution and its interested agents, whole cloth or in whateverway theychoose. altering the terrain on which the institution exists Rather,the very problems theyconfront and their and the pressures that generate agent incentives. sense of the range of the possible is shaped by the But agencytheorists have the virtue of contours of the broader institutional and political emphasizing the fact that agents create, design, and landscape in which theyare located. Referencing change institutions. In her language, actors have an idea of Pierson and Hacker,Thelen emphasizes politics,different conceptions of the way things that the background structure constrains agents by should be arranged. Contestation among agents providing them with a "menu of options from with differing politics can affect the spectrum of which [they] are forced to choose" (p. 287). possibilities for institutional transformation and the Change overtime in the background context in direction that transformation takes. The which agents act can, moreover, remove from structuralist perspective helps define the resources possibility strategies and goals that evenshortly available to actors; the agencyperspective focuses before were very much favored by actors. on who creates and reforms institutions and in Normally these twodifferent theoretical what way. perspectivesvieweach other with considerable This first move makes possible a second distrust: Structuralists think that agencytheorists move and that is to complicate the notion of believe that institutions are infinitely plastic and "feedback" that is deployed in historical that theyare the direct and desired result of agent institutionalism. Feedback, more specifically choice, when in fact choice is invariably highly "positive feedback", is typically a notion that is constrained, actors rarely have a clear idea of what associated with path dependencyand the theywant, and institutions, for manystructural reproduction of institutional systems, not with reasons, are deeply resilient. For their part, agency change. Positive feedback mechanisms create so- theorists think that structuralists overvalue called "increasing returns to power" or "situations constraint and underappreciate the creative role of in which victors at one stage impose institutional agents in the design and change of institutions. solutions that reflect and entrench their interests, Thelen thinks that all of this mutual mistrust is thus biasing outcomes in the next round" (p. 289). misplaced--or at least unnecessary.Infact she finds While not denying that such dynamics can be significant commonality between the Swenson found (she suggests that her own rendering of the perspective onthe one hand and that of the U.S. case has this character), Thelen draws on structuralists on the other: "these analyses all structuralist notions of change as well as her own highlight the need to situate the interpretation of notions of agencyand politics to suggest that specific choice points within a broader temporal positive feedback mechanisms can be destabilized,

-3- H-Net Reviews

creating the possibility for modification or reform conceptualized. Theyare simply twoways that of institutions evenastheyare being reproduced. come out of her analysis of the long history of the Citing the work of Paul Pierson, Thelen German vocational training system. The makes the point that institutions can fall into stimulating dimension of her theoretical work, processes of change because "changes in the however, isthat one is encouraged to think of broader social environment and/or the character of institutional change in her way and attempt to the actors themselves (among other things) can, come up with additional mechanisms for overtime, produce a significant and unintended understanding the gradual character of change in "gap" between the goals of designers and the way institutions. institutions operate" (p. 294). The existence of There are, however, a number of important such gaps (a structural effect) can lead to the questions about the theoretical movesthat Thelen destabilization of positive feedback processes by makes about which there needs to be clarity first, altering the alignment of interests that surround before we "go down her road." Some of the and support specific institutional designs. This, in questions have already been posed by some of her turn, unleashes "political" contestation overthe structuralist friends. Pierson, for example, finds the meaning, function, and design of the institution in notions of layering and conversion to be very society.Under such circumstances, she claims, the promising mechanisms for the conceptualization of out-of-line institutions do not necessarily get institutional change, but finds them also to be abandoned or destroyed. Instead, theycan be exceedingly underspecified.[1] Under what modified in various ways by coalitions of actors conditions will layering or conversion occur (as seeking to makethem reflect their interests. Any opposed to simple reproduction or to the collapse institutional system has several dimensions of or abandonment of the institution?). Thelen positive feedback and not all of those mechanisms generates the actors that convert and layer out of become targets for change. Hence portions of the the narrative flow ofher specific German case, but system can be reformed while other parts of it she does not provide general conditions for remain unchanged. The overall effect, however, determining when such actors will emerge, how can be quite significant for the institutional theycan be identified and whytheywill act to system--as the long durée account of the German reform the institution in the way that theydo. This vocational system attests. is a problem for the development of the notions in Thelen offers twodifferent forms of this general theoretical terms. kind of partial institutional transformation: This same area of under-specification also conversion and layering. By conversion, Thelen leavesThelen vulnerable to criticism from a refers to dynamics of change in which actors not constructivist, pragmatist point of view. originally associated with the construction or Constructivists with a pragmatic bent would begin design of the institutional system--indeed, even their criticism by focusing on her those who were excluded from the system when it conceptualization of agencyand its relationship to wasinitially designed ("losers" is her term)--find structure. Ultimately,Thelen conceivesofagents, an opportunity,due to structural destabilization, to such as "capitalists" or "large metalworking firms" takeoverthe institution and redirect its goals. This, as entities that are very well defined and whose she claims, is what happened to the original interests are givenbythe structural context in German artisan-controlled vocational training which theyare located. Agents and structure are system in the middle of the twentieth century as separate and typically the constitution of the the institution was appropriated and reformed by structure precedes the constitution of the agent. both industrial and trade union actors. Layering Structure imposes constraint on certain types of refers to modifications to institutional systems that action while at the same time enabling other sorts are added to an existing structure, as when trade of action. Crucially,for her,this does not union participation in monitoring and curricula unambiguously determine the way that agents will design was formally added to the German act because theyengage in "politics" with other vocational system in a 1969 law. interested/competing actors. The strategizing and It is not clear that Thelen wants us to believe compromising that such politics entails invariably that conversion and layering are the only ways in alters the outcomes achievedfrom that which which change within continuity can be would have been desired by anyactor individually.

-4- H-Net Reviews

This interest in contingencyislaudable, but it is all forms of action and relations, which also have a of a very modest kind that pays tremendous history and alternative social ties, constitute the deference to the power of structure overthe scope conditions out of which institutional recomposition of possibilities confronting actors. Once the actors and change becomes possible. Had Thelen have created a newarrangement, that newstructure conceivedofagencyand its relationship to clearly defines the interests of the actors and the structure in this way,she may have been able to array of possibilities for action that confront them. point to an array of potential developmental But this privileging of structure overseparate possibilities for training in Germanytoday,rather and independent agents is not the only way to than simply lament the potential collapse of a conceive ofthis relationship. It is also possible to century old system of practice. think of agents and structure being constituted How Institutions Evolve is thus not a perfect simultaneously in the process of action. Todothis, book. But it is so elegantly constructed and lucidly one needs to understand the imperativesimposed argued that it makes it possible to shift debate by structural designs to be ambiguous and the between structuralists, rational choice agency interests and evenidentities of actors to be ill- theorists, and constructivists on the character of defined and plastic, existing in a flowof change and the significance of temporality and communicative exchanges among contending choice in that process to a much higher and more actors regarding their possibilities for action and sophisticated level. the ends desired. In this kind of social and reflexive Note process, institutions and background structures become not constraints, but resources for the [1]. Paul Pierson, Politics in Time.History, creative reproduction of social identity and Institutions and Social Analysis (Princeton: collective practice. Through the creation or Press: 2004), pp. 139-142. modification of an institution, for example, actors define who theyare and howtheywould liketobe governed. Crucially,insodoing, theyare also aw are of the provisionality and contingencyof their creations. Actors knowthat institutional arrangements may not work or adequately solve the problems theywere designed to address. So theyinformally cultivate (retain) alternative identities and strategies for themselves at the same time that theyseek to modify and optimize their ownformal creations. So conceived, institutional change and institutional reproduction are permanently intertwined, as collective actors continually evaluate in the present the future viability of the habits and routines theyinherit from the past. Modification of practice and recomposition of institutions is a continuous process. The advantage in thinking of social recomposition in this historical and continuous wayasopposed to the rigidly delineated structure- action way that Thelen does is that it forces one to look at the formal systems of rule that guide action as multiple possibility creating, rather than as simply and unambiguously constraining. It also directs the eye to the informal, compensatory, experimental forms of action that all agents engage in, evenastheyengage in the reproduction of a particular institutional set of rules. Such alternative

-5- H-Net Reviews

Citation: Gary Herrigel. "ReviewofKathleen Thelen, How Institutions Evolve: The Political Economy of Skills in Germany,Britain, the United States and Japan, H-German, H-Net Reviews, . URL: http://www.h-net.msu.edu/reviews/showrev.cgi?path=150441145899120. Copyright © 2006 by H-Net, all rights reserved. H-Net permits the redistribution and reprinting of this work for nonprofit, educational purposes, with full and accurate attribution to the author,web location, date of publication, originating list, and H-Net: Humanities & Social Sciences Online. For anyother proposed use, contact the Reviews editorial [email protected] .

-6-