P1: KcS

0521837685pre Thelen 0521837685 June 19, 2004 10:36

How Institutions Evolve

THE POLITICAL ECONOMY

OF SKILLS IN GERMANY,
BRITAIN, THE UNITED
STATES, AND JAPAN

KATHLEEN THELEN
Northwestern University

v
P1: KcS
0521837685pre Thelen 0521837685 June 19, 2004 10:36

PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE

The Pitt Building, Trumpington Street, Cambridge, United Kingdom

CAMBRIDGE UNIVERSITY PRESS

The Edinburgh Building, Cambridge CB2 2RU, UK
40 West 20th Street, New York, NY 10011-4211, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
Ruiz de Alarcón 13, 28014 Madrid, Spain
Dock House, The Waterfront, Cape Town 8001, South Africa
http://www.cambridge.org

© Kathleen Thelen 2004

This book is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press.

First published 2004

Printed in the United States of America

Typeface Janson Text Roman 10/13 pt. System LATEX 2ε [TB]

A catalog record for this book is available from the British Library.

Library of Congress Cataloging in Publication data

Thelen, Kathleen Ann.
How institutions evolve : the political economy of skills in Germany, Britain,
the United States, and Japan / Kathleen Thelen.
p. cm. – (Cambridge studies in comparative politics)
Includes bibliographical references and index.
ISBN 0-521-83768-5 – ISBN 0-521-54674-5 (pbk.)
1. Employees – Training of – Europe – Case studies. 2. Employees – Training of –
United States – Case studies. 3. Employees – Training – Japan – Case studies.
4. Occupational training – Europe – Case studies. 5. Occupational training –
United States – Case studies. 6. Occupational training – Japan – Case studies.
I. Title. II. Series.
HF5549.5.T7T445 2004
331.25 92–dc22 200404785

ISBN 0 521 83768 5 hardback

ISBN 0 521 54674 5 paperback

vi
P1: KcS
0521837685pre Thelen 0521837685 June 19, 2004 10:36

Contents

Preface page xi

1 THE POLITICAL ECONOMY OF SKILLS IN

COMPARATIVE-HISTORICAL PERSPECTIVE 1
Skills and Skill Formation 8
The Argument in Brief 20
Theories of Institutional Genesis and Change 23
The Origins and Evolution of Institutions: Lessons
from the Present Study 31
Outline for the Book 37
2 THE EVOLUTION OF SKILL FORMATION
IN GERMANY 39
The Importance of the Artisanal Economy in the
Evolution of Skill Formation in Germany 42
Strategies of the Large Machine and Metalworking
Companies 55
Political Coalitions and the Evolution of the System 63
The Political Coalition against Reform 79
3 THE EVOLUTION OF SKILL FORMATION
IN BRITAIN 92
State Policy and the Fate of the British Artisanate 93
Union and Employer Strategies in the
Metalworking/Engineering Industry 104
Reform Efforts before World War I 118
The Impact of War and Its Aftermath 133
Comparisons and Conclusions 145

ix
P1: KcS
0521837685pre Thelen 0521837685 June 19, 2004 10:36

Contents

4 THE EVOLUTION OF SKILL FORMATION IN

JAPAN AND THE UNITED STATES 148
The Evolution of Skill Formation in Japan 149
The Role of the State and the Fate of the Japanese
Artisanate 151
Strategies of the Large Metalworking Companies 163
The Evolution of the Japanese Management System 166
Germany and Japan Compared 174
The Evolution of Skill Formation in the United States 177
Skill Formation in Early Industrial America 178
Union and Employer Strategies in the Metalworking
Industry before World War I 186
The Politics of Training during and after World War I 202
Comparisons and Conclusions 212
5 EVOLUTION AND CHANGE IN THE GERMAN
SYSTEM OF VOCATIONAL TRAINING 215
The Evolution of the System under National Socialism 219
Vocational Training in Postwar Germany 240
Contemporary Developments in the German Training
System: Erosion through Drift? 269
6 CONCLUSIONS, EMPIRICAL AND THEORETICAL 278
Cross-National Comparisons: The Origins of Divergent
Skill Regimes 278
Institutional Complementarities 285
Institutional Evolution and Change 292

Bibliography 297
Index 323

x
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in

Comparative-Historical Perspective

The past two decades have witnessed an enormous outpouring of litera-

ture on the putative effects of “globalization” on the political economies
of the advanced industrial countries. A good deal of this literature was in-
spired by early, sometimes rather breathless predictions of a trend toward
convergence in the institutional arrangements governing these political
economies. Such convergence, it was argued by some, would result from
the pressures imposed by footloose firms engaged in “regime shopping”
which would in turn drive competitive deregulation among the advanced
countries (see, for example, Kapstein 1996; Kurzer 1993).1 These prospects
were especially worrisome to students of Europe’s “corporatist” political
economies, which had long been admired as models of economic efficiency
and social equality.
In the meantime, however, a good deal of evidence has accumulated that
calls into question arguments about a convergence among the institutional
arrangements that characterize different political economies (Berger and
Dore 1996; Brown, Green, and Lauder 2001; Ferner and Hyman 1998;
Garrett 1998; Iversen, Pontusson, and Soskice 2000; Kitschelt et al. 1999;
Streeck and Yamamura 2002; Vogel 2001; Wallerstein and Golden 1997).
Although there are certainly changes afoot in all countries, a number of
scholars have pointed to systematic and apparently enduring differences in
the organization of capitalism across the advanced industrial countries. Dif-
ferent authors characterize these differences each in his or her own way, but
the consensus that has emerged is truly striking. Boyer and Hollingsworth
write of distinctive national “production regimes” which are defined by a

1 Harry Katz and Owen Darbishire make a more nuanced argument about the pace and scope
of common trends (Katz and Darbishire 1999; see also Martin and Ross 1999).

1
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

set a mutually reinforcing institutional arrangements that together support

different types of firm strategies in international markets – situating the
United States and the United Kingdom at one end of a continuum and
Germany and Japan at the other (Boyer and Hollingsworth 1997; see also
King and Wood 1999).2 This work resonates with Streeck’s analysis, which
distinguishes between “liberal” market economies such as the United States
and Britain and “socially embedded” political economies such as Japan and
Germany (Streeck 2001). Michel Albert’s popular book draws a distinc-
tion between what he calls the “Anglo Saxon” and the “Rhineland” ver-
sions of capitalism, which, as Streeck has pointed out, also situates “not
just Stockholm but also Tokyo” on the banks of the Rhine (Albert 1993).
All this, in turn, maps well onto the analysis of David Soskice, who has
characterized cross-national differences in advanced capitalism in terms of
a broad distinction between “coordinated” and “non-coordinated” market
economies, and more recently, in collaboration with Peter Hall, of “coordi-
nated” versus “liberal” market economies (Hall and Soskice 2001; Soskice
1991).
What all these works share in common is a perspective on national mod-
els of capitalism that are characterized by distinctive institutional arrange-
ments which in turn support specific kinds of strategies on the part of firms
in international markets. These authors all point to more or less the same
set of institutional arrangements that have traditionally been defined as cen-
tral to the functioning of these political economies – financial institutions,
industrial relations institutions, vocational training systems, bank–industry
links, and more recently, welfare state institutions and policies (Ebbinghaus
and Manow 2001; Estevez-Abe, Iversen, and Soskice 2001). For all their dif-
ferences, these authors make very similar distinctions across countries, and
draw an especially sharp line between “organized” (or “embedded”) capi-
talist economies such as Germany and Japan on the one hand, and “liberal”
market economies such as the United States and the United Kingdom on
the other.3 The former are characterized by “patient capital,” coordinated

2 The literature on the political economy of advanced capitalism from the 1970s and 1980s
similarly focused on distinctive national models (for example, Hall 1986, Zysman 1983, and
corporatism theorists of the 1970s). The newer literature on “varieties of capitalism” is an
outgrowth of that earlier work, though with some important theoretical innovations. For
an extended discussion, see Thelen (2002b).
3 In fact, all these various categorization schemes also have trouble sorting the same set of
“intermediate” or hard to classify countries, including France and Italy.

2
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

employers, and various forms of labor-management cooperation. The lat-

ter, by contrast, are defined by short-term financing arrangements, frag-
mented employers, and more adversarial industrial relations systems.
What these characterizations also share is a view of these systems as
more or less integrated wholes, in which the various parts work together
in ways that are mutually reinforcing. But what brought these systems into
being? And what holds them together today? And, since no one would argue
that these systems are not changing in fundamental ways, how should we
characterize the dynamics of change and weigh these against the forces of
institutional reproduction? These questions are at the top of the agenda in
the context of current globalization pressures, and yet on these issues the
consensus begins to break down.
Hall and Soskice’s characterization of the ideal-typical “organized” and
“liberal” market economies emphasizes tight coupling and multiple func-
tional interconnections among the various institutional arrangements that
make up a national economic “system” (Hall and Soskice 2001). For these
authors, the existence of institutional complementarities across various are-
nas contributes to the robustness of the system as a whole, for two rea-
sons. One is that institutional frameworks provide the foundation on which
(nationally distinctive) competitive advantage rests, so that key actors (espe-
cially employers) who have organized their strategies around these institu-
tions will be loath to part with them. The other is that, even if they wanted
to change some aspects of the system, shifting rules in one arena would
require adjustments in other, neighboring realms, which both increases the
costs of change for them and also multiplies the political obstacles to such
change. That said, however, and despite the fact that Hall and Soskice over-
all see the systems as very robust, tight coupling and strong institutional
complementarities also seem to suggest that a major disruption in one realm
(for example, financial institutions) would immediately “radiate” and trans-
late into significant strains and change in neighboring realms (for example,
collective bargaining).
Streeck’s historical–sociological account also stresses institutional com-
plementarities but it downplays the functional and economic logic in favor
of a more singularly political view. Streeck views national “systems” as the
product of past and ongoing political intervention and tinkering, of active
maintenance and re-setting. Whatever structural or functional coherence
we may now observe in these systems, he argues, “had to be continuously es-
tablished, restored, redefined and defended against all sorts of disorganizing

3
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

forces” (Streeck 2001: 31). He emphasizes how national models are not the
product of a grand design, and “Ex post accommodation . . . seems to have
been at least as important . . . as a priori calculations of the advantages of
compatibility and complementarity under conditions of interdependence”
(Streeck 2001: 31). Unlike Hall and Soskice’s equilibrium-based view of
institutions, Streeck sees these arrangements as inherently less coherent
and therefore also not at all self-equilibrating, which leads to a picture that
is overall less sanguine about the robustness of such systems. At the same
time, however, and for many of the same reasons, this alternative view would
not necessarily see such systems as vulnerable to the kind of “unraveling”
mentioned above in the event of strong perturbations in one realm.
The premise of this book is that in order to understand the likely fu-
ture of the institutions that make up the different “varieties of capitalism”
we need a better sense of where these institutions came from, what has
sustained them, and what are the ways in which they have changed over
time. This is not the first time these systems have experienced strain, and
understanding how they evolved in the past can yield new insights into the
modes and mechanisms of change through which they continue to develop
today. My analysis focuses on the institutions of skill formation because
they constitute a key element in the institutional constellations identified
by all the authors cited above. In fact, one recent strand of scholarship sees
skills and skill formation systems as causally central to the development and
articulation of social policy preferences generally, and thus foundational for
the development and maintenance of different systems of social protection
across the developed democracies (see, especially, Iversen and Soskice 2001;
also Iversen 2003).
I approach the politics of skill formation from two angles, pursuing both
a cross-national and a longitudinal dimension. First, the cross-national com-
ponent of the study traces the origins of different skill formation regimes,
focusing especially on Britain and Germany, and with only slightly less de-
tailed treatments as well of Japan and the United States. This part of the
book asks the questions: Why did different countries pursue such different
trajectories in terms of plant-based training? And: how did the evolution of
training institutions interact with the development of “collateral” organiza-
tions and institutions – especially labor unions and employer associations,
and industrial relations institutions?
One of the most widely cited differences between “coordinated” market
economies (such as Germany and Japan) and “liberal” market economies
(such as the United States and United Kingdom) is that the former
4
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

support more and better plant-based training.4 My comparative historical-

institutional analysis traces the origins of these contemporary differences
back to the nineteenth century. Contemporary differences in skill forma-
tion go back to important differences in the character of the settlement be-
tween employers in skill-based industries, artisans, and early trade unions. I
show how the development of skill formation in the early industrial period
interacted with the development of collective bargaining institutions and
nascent labor unions and employer organizations in ways that set coun-
tries on different national trajectories. The present analysis identifies some
similarities between Germany and Japan (both as “coordinated” market
economies), and between Britain and the United States (as “liberal” mar-
ket economies). However, it also underscores the substantial differences
between these cases in how training is organized – with enormous implica-
tions for the type of skill formation that each country institutionalized and
for labor’s role within it.5
The cross-national component of the analysis draws on an excellent
secondary literature on the histories of unions, employer organizations,
and skill formation, as well as on primary documents where these were
needed to address the specific questions around which this research was
organized. The main goal is to situate a number of country experiences

4 Mine is not the first study to draw attention to these particular cases. As mentioned above,
the different literatures on varieties of capitalism tend to focus on the same basic clusters
of countries. Likewise, in the literature on skills in particular, it is not uncommon to draw
a broad distinction between Germany and Japan on one hand (as “high-skill” economies)
and the United States and United Kingdom on the other (as “low-skill” or strongly skill
bifurcated economies) (for example, Ashton and Green 1996; Brown, Green, and Lauder
2001). This convention follows pioneering work by Finegold and Soskice (1988), who draw
a sharp line between “high-skill” and “low-skill” equilibrium countries, though in the mean-
time this somewhat simplified distinction has been superseded by a more differentiated view
that emphasizes the different types and mixes of skills produced within any one country.
This is also why I adopt the somewhat less normatively inflected language of “coordinated”
versus “liberal” training regimes, which are characterized by different mixes of strengths
and weaknesses in skill formation. See the discussion below.
5 Ashton and Green (1996) propose an explanation of these outcomes that is broadly com-
patible with mine. However their historical analysis is very fleeting (five to ten pages per
country, and in each case covering both plant-based and school-based training). Their ex-
planation of these outcomes hinges on the behavior of highly aggregated actors (“the ruling
class,” the “bourgeoisie,” the “aristocracy”), and the conclusions – although not wrong –
are highly simplified. Many of their assertions (for example, the idea that the Handwerk
sector was “undermined” by industrialization or the implication that unions pushed for
skill standardization against employer opposition) are not consistent with the results of the
historical–empirical research presented here (Ashton and Green 1996: 142–3).

5
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

within a theoretical framework that can illuminate causal mechanisms at

work across a number of cases. Each of the countries included in the broader
study can be – has been – characterized as unique in the literature on politi-
cal economy and labor development – Britain as the first industrializer; the
German “Sonderweg”; American and Japanese “exceptionalism.” Without
taking away from the fundamental uniqueness of each case, this project
attempts to put their experiences with skill formation and labor incorpora-
tion side by side to shed light on systematic parallels and differences among
them.
Second, the longitudinal dimension of the research tracks the development
of the German vocational training system over a longer time frame than
the other cases – indeed, up to the present. This aspect of the study tackles
the question of how institutions evolve and is organized around a somewhat
different puzzle. Germany’s vocational training system has been held up as
an exemplary solution to a number of knotty coordination problems that
plague most private sector training regimes. The German system encour-
ages firms to invest in worker skills, and it provides mechanisms to assure
that apprentices will receive high quality training. More generally, voca-
tional training institutions in Germany are typically seen as part of a larger
institutional package which, along with centralized collective bargaining,
strong bank–industry links, and encompassing employer associations and
labor unions, are seen as underpinning the country’s high skill, high wage,
high value-added (“high everything”) economy. The vocational training
system is viewed as a key element in a larger institutional complex that ac-
tively supports a production regime organized around a kind of “diversified
quality production” that reconciles Germany’s comparatively strong unions
with strong performance in world manufacturing markets (Streeck 1991).
Scholars coming out of different theoretical perspectives have offered
different views of the German system. From a functional–utilitarian per-
spective, German vocational training institutions have been seen as part of a
complex institutional configuration that, among other things, supports em-
ployer coordination around a “high skill equilibrium” (Soskice 1991). From
a power–resource perspective, the German vocational training system has
been assumed to be a reflection of working class strength (for example,
Gillingham 1985). From a sociological–cultural perspective, these institu-
tions have been seen as a prime example of a more general, and distinctively
German, mode of self-governance, which operates through the country’s
social partners and without much direct intervention from the state (see,
for example, Lehmbruch 2001).
6
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

There is a way in which, looking at these institutions today, all these

characterizations contain an element of truth. After an examination of the
historical evolution of this system, however, they all seem rather puzzling.
The core institutional innovation around which the German system came
to be built was legislation passed in 1897 by an authoritarian government
motivated politically by a desire to shore up and support a conservative
class of independent artisanal or handicraft producers that could serve as
a bulwark against the surging and radical working class movement. The
1897 legislation proved to be very important to stabilizing a certain type
of plant-based apprenticeship that in other countries was rapidly deteri-
orating into cheap child labor. Against prominent utilitarian-functionalist
characterizations focusing on current effects, however, these institutions
were not originally designed with the economic interests of the industrial
sector in mind (industry was excluded in fact) and they were certainly not
meant to reconcile strong unions with anything. Against perspectives that
assume that these institutions must have been the creation of unions, we
find that organized labor played no role in promoting the original legisla-
tion, and in fact, the Social Democratic party opposed it. Finally, against
the cultural–sociological perspective mentioned above, it becomes clear
that the particular type of social partnership of which these institutions are
(now) seen to be a part was really nowhere on the horizon.
How did we get from there to here? Not through a wholesale breakdown
of the old institutions and their replacement with new ones. One of the
striking features of the system is the resiliency of core elements even in the
face of enormous disruptions over the twentieth century, which of course in
Germany include several regime changes, the incorporation of the working
class, defeat in not one but two world wars, occupation, and transitions
both into and out of fascism. Although changes certainly occurred at these
junctures, what is remarkable and in need of explanation are some striking
continuities in key features of this system despite these disjunctures. This
case, in short, calls for an analysis both of the mechanisms of reproduction
that sustained these institutions and the mechanisms behind their functional
and distributional transformation over time.
An examination of the evolution of German vocational training institu-
tions can provide a window on larger questions of institutional evolution
and change. I argue that this case illustrates the way in which institutional
reproduction is often inextricably linked to elements of institutional trans-
formation of the sort that brings institutions inherited from the past into
synch with changes in the political and economic context. In the case of
7
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

German vocational training institutions, I analyze how the emergence and

incorporation of new groups whose participation was not anticipated at
the time the system was created drove the development of these institu-
tions forward in ways that, over time, produced a system whose political–
distributional effects are completely different from those envisioned by the
original designers.
While emphasizing and explaining the disjuncture between original in-
tent and ultimate effects, the analysis of the evolution of German vocational
training institutions also speaks to the issue of institutional reproduction –
a more problematic concept than typically recognized (for a discussion,
see Thelen 1999). A point of departure for my analysis is that we need to
pay more attention to how institutions created at critical junctures in the
sometimes quite distant past actually make it to the present, given the mag-
nitude of some of the intervening events and developments. I argue that it
is not sufficient to view institutions as frozen residue of critical junctures,
or even as “locked in” in the straightforward sense that path dependence
arguments adapted from the economics literature often suggest. In politics,
institutional reproduction can be partly understood in terms of the increas-
ing returns effects to which this literature has drawn our attention – but
only partly. Only partly because as one observes the development of the
German system over a long stretch of time, it becomes clear that institu-
tional survival often involves active political renegotiation and heavy doses
of institutional adaptation, in order to bring institutions inherited from the
past into line with changes in the social and political context. The question
of the mechanisms of institutional reproduction and institutional change is
thus the second major theme explored in the pages below.

Skills and Skill Formation

Vocational training institutions occupy a central role in most characteriza-
tions of the various political–economic systems cited above – and for good
reason. Skills are associated with a variety of outcomes of interest to polit-
ical economists. The acquisition of skills and investment in human capital
are seen by many economists as “an engine of growth” (Acemoglu and
Pischke 1999a: F112), and deemed to be “absolutely central to countries’
growth performance” (Booth and Snower 1996: 1). Several studies point
to a strong link between skills and productivity (Acemoglu 1996; Black
and Lynch 1996; Bishop 1994; Lynch 1994: 21–2). In addition, growth
and human capital development also figure prominently in “endogenous
8
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

Skills and Skill Formation

growth theory,” which argues that a country’s knowledge base is an impor-

tant resource for innovation and which has linked cross-national differences
in education to persistent disparities in per capita income across national
economies (Romer 1990).
Political scientists and sociologists are as interested in the social conse-
quences of various skill development systems as they are in the economic im-
pact (for example, Brown, Green, and Lauder 2001). As mentioned above,
analysts have begun pointing to the consequences of different skill forma-
tion regimes for broad social and institutional outcomes, including gender
inequality and divergent social policy regimes (Estevez-Abe, Iversen, and
Soskice 2001; Mares 2000; Iversen and Soskice 2001; Lauder 2001). Differ-
ent training systems have also been linked to a range of more specific politi-
cal and social effects. For example, training regimes in what Hall and Soskice
call “liberal market economies” tend to be associated with rather wide gaps
in the opportunities available to different types of workers, with implica-
tions for, among other things, income inequality (also Crouch, Finegold,
and Sako 1999: 3). In general, less skilled workers have more opportu-
nities for advancement in what Hall and Soskice call “organized market
economies” which are, in turn, also characterized by greater income and
wage equality.6 Wage inequalities based on skill have been rising as a result
of changing technology, which increases the skill premium as well as the
overall skill intensity of production (Levy and Murnane 1992; Katz and
Murphy 1992; Gottschalk and Smeeding 1997).
Sustained attention to vocational training institutions in the literature
on the political economy of the advanced industrial countries can be traced
back largely to Streeck’s pioneering work in this area (Streeck 1992b). Based
on an analysis of the German case, Streeck showed how strong private
sector training within the context of a standardized and uniform national
system for vocational education supported firm strategies based on what
he called diversified quality production (Streeck 1991). For Streeck, voca-
tional skills (broad-based and widely available) constitute a crucial compet-
itive advantage for firms operating in “technologically and economically
volatile markets” (1992b: 166). But despite the benefits of training, Streeck
noted, firms operating according to strict economic rationality will system-
atically under-invest in the kind of broad and transferable skills necessary

6 For example, Acemoglu and Pischke note that low education entry-level workers are more
likely to get training in Germany than in the United States (Acemoglu and Pischke 1999a:
F129).

9
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

to compete in these markets. As he puts it: “firms that create only those
skills that they need may well end up with less than they need. Cost- and
profit-consciousness are more part of the problem than of the solution”
(Streeck 1992b: 17). In Streeck’s view, cultural or political constraints that
bend the logic of individual market incentives – in this case encouraging
firms to “over invest” in skills – may turn out “to be more conducive to
economic performance under given technological and market conditions
than others” (Streeck 1992b: vii).
Given this posing of the problem, Germany was long considered a
model of successful vocational training and skill formation for the ad-
vanced industrial countries generally. The comparisons to other countries
were mostly invidious (see, for example, Berg 1994; Oulton and Steedman
1994; Finegold 1993). In one important contribution, for example, David
Finegold and David Soskice examined the institutional bases of Germany’s
“high skill” and the United Kingdom’s “low skill” equilibria (Finegold and
Soskice 1988). In particular, they argued that Britain’s chronic undersupply
of training went back to public good or free rider problems, and the re-
sulting dearth of skills in the economy encouraged firms to pursue product
strategies premised on low skills, which in turn discouraged investment in
skills, and so on. This dynamic provided the backdrop to their analysis of
the inability and/or unwillingness of the British government to take action
to break the cycle (Finegold and Soskice 1988: 25ff).
In the meantime, the literature has begun to paint a somewhat more dif-
ferentiated pattern of strengths and weaknesses (see, for example, Crouch,
Finegold, and Sako 1999; Culpepper 2003; Culpepper and Finegold 1999;
Hall and Soskice 2001; Green and Sakamoto 2001). Thus, for example,
Finegold’s more recent work highlights Germany’s continued high invest-
ment in initial manufacturing skills (apprenticeship) but notes deficits in
the commitment of firms to further training – now arguably more impor-
tant than ever in the context of rapidly changing production technology –
(see Pichler 1993), as well as the overall scarcity of certain high-end
(information technology and engineering) skills (Crouch, Finegold, and
Sako 1999; Atkins 2000). Conversely, the United States, long viewed as a
skills “laggard,” is now getting more credit for producing an abundance
of high-end skills – having become a net exporter of skills in informa-
tion technologies, for example – despite continued worries about the coun-
try’s under-investment in traditional manufacturing skills (Hall and Soskice
2001; Smith 2000). In a way, the emphasis in the political economy litera-
ture has shifted from an effort to identify overall differences in the quantity
10
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

Skills and Skill Formation

of training cross-nationally,7 toward a more fine-grained analysis of cross-

national differences in the particular mix of jobs and qualifications that char-
acterize different political economies (Estevez-Abe, Iversen, and Soskice
2001; Crouch, Finegold, and Sako 1999).

Economists on Skills
As a first step in situating this literature and in mapping the empirical cases, it
is useful to summarize briefly the way in which economists have approached
the issue of skill formation. A point of departure for the political economy
literature cited above is training market failures going back to poaching
externalities, a view which for most of the first half of the twentieth cen-
tury was widely accepted among economists (see, for example, Pigou 1912;
Stevens 1996). The crux of the problem was seen to consist of a collective
action dilemma. Specifically, firms that need skills face a choice: whether to
provide their own training (at some cost to themselves), or attempt to secure
skilled labor on the labor market (in effect, to attract workers from other
firms that have invested in training). Since firms are themselves competitors
in labor and product markets, each individual company will be tempted to
poach workers from other firms, thus avoiding the costs associated with skill
formation and in fact making off with the investment of its competitors.
The more firms that choose this strategy, the greater will be the costs to
those firms that do train; they incur both the costs of training itself and the
costs of competing with non-training firms which – free of these costs –
can offer these workers a wage premium. If all firms pursue non-training
strategies, then of course all are worse off since this would diminish the
overall stock of skills on which all rely.
In a widely cited contribution, Gary Becker called into question this pos-
ing of the problem (1964). Becker distinguished between general skills and
specific skills and argued that poaching externalities would not be a prob-
lem in either case. For general skills – defined as those skills that are fully
transportable and hold value to many employers – it is quite true that firms
have no incentive to invest in training. But even if firms do not invest, work-
ers will, or in Becker’s words, “it is the trainees, not the firms who would
bear the cost of general training and profit from the return” (1993: 34).8

7 The available, mostly OECD, data are anyway notorious for their lack of comparability (for
example, OECD 1998: 10–11). For a discussion see also Lynch (1994).
8 Note that this formulation does not address the question of the supply of training, however.

11
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

Assuming perfectly competitive labor markets, skilled workers who possess

general skills will be paid a wage equal to their marginal product, which
means that even though the firm has no incentive to pay for training (be-
ing unable to capture returns on that investment), the workers themselves
would do so (in the interest of their future higher wages). Becker’s argument
thus turns the poaching issue on its head: In a completely competitive labor
market, poaching is not a cause of under-investment in general skills, but
rather a mechanism assuring that a worker’s skill is valued at full marginal
product, thus providing an incentive for the workers themselves to acquire
training.
Specific skills, according to Becker, are completely non-transportable and
have value only to the particular firm in which the worker is employed
(Becker 1993: 40). Poaching again is not a problem, in this case because
firm-specific skills by definition are valuable only to the present employer.
Here the equilibrium outcome is for the firm and the worker to share the
costs of training. Facing no external competitors for the skills they are
imparting, firms can pay workers a wage below marginal product in order
to realize a return on the training investment. Workers will be willing to
share the costs since the resulting wage will lie above the external market
wage but below marginal productivity wage after training (Becker 1993: 42).
The problems that Becker identifies as sources of possible market fail-
ure resulting in under investment in skills go back not, as in the previous
literature, to poaching externalities, but rather mostly to capital market con-
straints and especially credit constraints on the trainees. Although willing to
invest in their own general training, trainees may find it hard to secure credit
to do so because of the high risk of default on a loan and/or the uncertainty
of the return on training (Becker 1993: 39–40 fn12). As Stevens (1999: 20)
points out, the expected returns on training are the highest for workers at
the beginning of their careers, precisely the time at which their accumulated
resources are lowest (also Acemoglu 1996; Acemoglu and Pischke 1999a).
Public policies that ease trainees’ credit constraints – for example, loans
or subsidies for training – could alleviate the sources of market failure iden-
tified by Becker. The logic of the argument suggests, however, that any
mechanism that effectively helps the worker defer or pay the costs of train-
ing will do. In traditional apprenticeship in many countries, for example,
fees paid by an apprentice’s family to the training firm and/or long periods
of apprentice indenture (in which firms continued to pay apprentices very
low wages even after they had acquired substantial skills) serve such a func-
tion. The logic of the argument points additionally to the benefits of some
12
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

Skills and Skill Formation

regulatory framework or system of accreditation that would ensure quality

training and therefore reduce the uncertainty of return on the trainee’s side.
As we will see, the decline of the British system of apprenticeship training
(for example) can be traced back to the failure to solve problems of uncertain
returns for either employers or trainees.

Beyond Becker
In the meantime a number of studies have emerged which take us “be-
yond Becker” (Acemoglu and Pischke 1999a) but also, paradoxically, back
to some of the training–market problems identified by the earlier literature.
This work responds in part to empirical anomalies that point to sources
of under investment in skills that appear not to be captured by the logic
of Becker’s model (Acemoglu and Pischke 1999a; Finegold and Soskice
1988). Most of this more recent scholarship does not take issue directly
with Becker’s analysis of the two ideal-typical models of training markets
characterized by perfect competition (general skills) and non-competitive
labor markets (specific skills), but research has focused on the provision of
training in the intermediate case of imperfectly competitive labor markets
for certain kinds of skills. Specifically, analysts have reacted to two aspects
of Becker’s argument: (1) the sharp distinction he draws between general
and specific skills (Estevez-Abe, Iversen, and Soskice 2001; Stevens 1996;
Stevens 1999),9 and (2) his implicit assumption that the market for general
skills will always be perfectly competitive (Acemoglu and Pischke 1998;
Acemoglu and Pischke 1999a; Acemoglu and Pischke 1999b).
First, Margaret Stevens attacks the problem by questioning Becker’s stark
categorization of skills as either general or specific. She focuses on an inter-
mediate category of skills, what she calls transferable skills.10 Such skills “are
of value to more than one firm, and there is competition between firms to
employ the worker, but competition is not sufficiently fierce that the wage
is driven up to the marginal product” (Stevens 1999: 19). Stevens and others
imply that very many skills fall into some kind of intermediate category.
In a similar vein, Finegold and Soskice note that a given firm may need
a particular mix of skills, and that while each one may be general the mix

9 Although Becker does recognize that most training is neither purely general nor purely
specific (Becker 1993: 40).
10 This seems similar to the category “industry skills” identified by Estevez-Abe, Iversen, and
Soskice (2001).

13
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

itself is specific (Finegold and Soskice 1988; Acemoglu and Pischke 1999a:
F124). Moreover, Franz and Soskice (1995) argue that the acquisition of
general and specific skills is complementary, that is, teaching firm-specific
skills reduces the costs of teaching general skills and vice versa.
The introduction of this intermediate category of transferable skills
(portable but not really “general” in Becker’s sense) re-introduces the prob-
lem of poaching externalities since in an imperfectly competitive labor mar-
ket the value of the skill (= wage) will not necessarily be driven up fully to
marginal product, and some of the training accrues to the firm that employs
the skilled worker (Stevens 1999: 20). This situation reduces the employee’s
incentive to acquire skills (since he will not be able to capture the full re-
turn on his training investment in the form of wages), but for the very same
reason, it also increases the employer’s incentive to assume a part of the
costs of training. But this, in turn, raises again the possibility that one firm’s
investment in training might accrue to some other firm – that is, a poach-
ing externality – leading to under investment in skills of this type (Stevens
1999: 27).
Acemoglu and Pischke argue similarly for the possibility of imper-
fect labor market competition and associated implications for training
(Acemoglu and Pischke 1999a: F127). Whereas Stevens and others (for
example, Estevez-Abe, Iversen, and Soskice 2001) attack the Becker the-
sis by drawing out more fine-grained distinctions among types of skills,
however, Acemoglu and Pischke focus instead on the structure of the labor
market. Against Becker’s thesis that firms will never pay for general training,
they note empirical examples (German apprenticeship is one) where firms
do bear a significant fraction of the cost of general training (Acemoglu and
Pischke 1999a: F113–4). By way of explanation they cite labor market im-
perfections that prevent skilled workers from claiming wages equal to their
full marginal productivity and that thus allow employers to earn rents on
their training (Acemoglu and Pischke 1999a: F120; Acemoglu and Pischke
1998: 80).
Acemoglu and Pischke consider various sources of labor market imper-
fections. Institutions or situations that result in low labor turnover, for
example, distort the competitiveness of labor markets, making it possible
for employers to retain a worker after training at a wage lower than marginal
product. One source of low labor turnover is monopsony power exercised
by large firms that may dominate a local economy. Firms in such a situation
could pay a wage above the local rate but still below the marginal produc-
tivity of its trained workers (Acemoglu and Pischke 1998: 80–1). Japanese
14
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

Skills and Skill Formation

training is frequently coded as involving “specific” skills in Becker’s sense,

but as the historical analysis below shows, it may be more properly un-
derstood as a case in which imperfectly competitive labor markets stem-
ming from monopsony power (and coordination among monopsonists) al-
low firms to invest in transferable skills.11
Another source of labor market imperfections that figures prominently
in Acemoglu’s and Pischke’s analysis involves labor market institutions that
compress wages, thus holding back skilled workers’ wages. Wage compres-
sion encourages firms to invest in general training, because, by holding
wages below marginal productivity, it again allows the firm to capture some
of the returns from the investment. This is at odds with a core assump-
tion in Becker’s theory, that wages for workers possessing general skills will
grow at the same rate as their marginal productivity. Acemoglu and Pischke
point out, however, that labor market institutions may prevent this, with
implications for firm investment in training: “With competitive labor mar-
kets, firms never pay for investments in general training, whereas when
labor markets are imperfect, firm-sponsored training arises as an equilib-
rium phenomenon” (Acemoglu and Pischke 1999a: F112).12 As we will
see below, union policies in Germany that encouraged wage compression
in the Weimar years were indeed associated with substantial advances in
firm-based training.13

Back to Politics
The analytic move made by Acemoglu and Pischke – from the assumption
of perfectly competitive labor markets to the observation that the degree

11 Blinder and Krueger (1996) show that labor turnover in Japan is less than half the U.S.
level. Japanese management policies, including seniority-based wages, company-level so-
cial policy, and reluctance to recruit experienced workers from other firms, contribute to
this outcome.
12 Acemoglu and Pischke also note other sources of labor market imperfections that have
similar effects to low labor turnover and wage compression. These include matching and
search frictions, which make it difficult for workers to quit their jobs and find new ones, and
asymmetric information about how much training workers have (that is, a firm which trains
its workers has an advantage over competitors because it has better information about the
workers’ skill levels) (Acemoglu and Pischke 1999a).
13 Additionally, the typical argument about compression is that this creates incentives for em-
ployers to replace relatively expensive unskilled workers with machines. Together with the
skills argument I am making, this points toward a high-skill, higher technology trajectory
(supporting what Streeck has called “diversified quality production”) (see also Moene and
Wallerstein 1995; Streeck 1991).

15
P1: JzG/JzK
0521837685c01 Thelen 0521837685 May 31, 2004 17:5

The Political Economy of Skills in Comparative-Historical Perspective

of competitiveness in the labor market can vary (with implications for the
incentives facing both firms and trainees) – is crucial to the present anal-
ysis because, as the case studies elaborated below show, training regimes
developed historically in tandem and in interaction with the development
of other labor-market institutions and organizations. In Britain, as we will
see, union structures and collective bargaining institutions developed in
ways that intensified and in some ways distorted competition among em-
ployers in both skilled labor and product markets. By contrast, in Germany
and Japan, developments with respect to union structures and collective
bargaining introduced labor-market imperfections that provided some in-
centives for firms to train while at the same time (re-)introducing collective
action problems associated with poaching.14
The point is that the structure and operation of skilled labor markets
can be significantly influenced by politics, and historically, the interaction
of the development of training regimes and of labor-market institutions
had a profound effect on what kind of skilled labor market employers (and
trainees) faced, as well as the kinds of solutions available for redressing
the particular (different) market failures that emerged in different contexts.
The problems (and solutions) that emerged historically are what lie behind
some of the striking contemporary national differences in training regimes
(Acemoglu and Pischke 1999a: F132).
In terms of outcomes, Peter Hall and David Soskice note that in lib-
eral market economies, the incentives are for young people to acquire skills
that are generally marketable rather than firm- or even industry-specific.
Companies may upgrade this education with some company training but
typically attempt to add only non-transferable (firm-specific) skills whose
full benefit they and only they can recoup (Hall and Soskice 2001). The
U.S. training regime thus does not favor strong firm-based investment in
private sector vocational training. It appears, however, to support very well
the production of a plentiful supply of “high-end” skills – for example, en-
gineering and programming – that thrive in a context that rewards strong
general (especially university) education and where demand for training on
the part of young people is driven by intense competition among firms.

14 Acemoglu and Pischke do not do the historical analysis, but they do stress the complemen-
tarities between training systems and labor market regulations (1999a: F136) and suggest
that “the more frictional and regulated labor markets [in Europe and Japan] may encourage
more firm-sponsored training” (Acemoglu and Pischke 1999b: 567). See also Peter Hall
and David Soskice who discuss institutional complementarities between training systems
and other political–economic institutions (Hall and Soskice 2001).

16