Methodological Pitfalls of Convergence Analysis
| Author | Katharina Holzinger |
| Published date | 01 June 2006 |
| Date | 01 June 2006 |
| DOI | http://doi.org/10.1177/1465116506063721 |
| Subject Matter | Articles |
Forum Section
Methodological Pitfalls of
Convergence Analysis
Katharina Holzinger
Hamburg University, Germany
271
European Union Politics
DOI: 10.1177/1465116506063721
Volume 7 (2): 271–287
Copyright© 2006
SAGE Publications
London, Thousand Oaks CA,
New Delhi
The study of cross-national policy convergence has become highly popular
in political science. The academic popularity of the topic significantly
increased in the 1990s. There is an ever-growing body of research that inves-
tigates the occurrence and the underlying driving forces of policy conver-
gence. Notwithstanding these efforts, we still have a limited understanding
of the extent and causes of policy convergence. Both conceptual and methodo-
logical heterogeneity impose important restrictions on the comparability of
the empirical findings.
Whether a study finds convergence, divergence or persistence of policies
depends very much on the measurement concepts used. As an abstract state-
ment, this seems trivial. However, Seeliger (1996) has already pointed to the
fact that, in empirical convergence analyses, little attention is paid to these
problems. This contribution presents a selection of the many methodological
problems one is confronted with in convergence analysis. It focuses on the
convergence of policies. However, similar problems may arise with respect to
the convergence of institutions, culture, habits, etc. Furthermore, this contri-
bution focuses on problems of measurement, leaving aside problems of explain-
ing convergence. The problems will be illustrated using data from the research
project ‘Environmental Policy Convergence in Europe’ (ENVIPOLCON).1
Problems of measurement of similarity
The measurement of policy convergence faces a number of problems, relating
both to ‘policy’ and to ‘convergence’. Usually, policy convergence is broadly
understood as an increase in policy similarity over time. This raises two ques-
tions. First, how can the ‘similarity’ of policies be measured? Second, how is
its ‘increase over time’ conceptualized? I start with problems of measuring
the similarity of policies before coming back to problems related to measur-
ing convergence.
Comparing policies
If we talk about cross-national policy convergence, we are usually referring
to a certain policy field or to a single policy measure in a certain policy area.
This policy (field) potentially exists in several countries, and the question is
whether it has become more similar in the observed countries and time
period. However, policy fields and single policies are highly complex subjects,
encompassing many dimensions that can be compared in order to assess their
similarity. In the literature, we generally find a broad list of policy dimen-
sions on which convergence might occur, including, for instance, policy
output or content, policy style or process, as well as policy outcomes (Bennett,
1991: 218). Another differentiation is suggested by Dolowitz and Marsh (1996:
349–50), which includes ‘policy goals, structure and content; policy instru-
ments or administrative techniques’ (see also Heichel et al., 2005: 831).
With respect to methodological questions, one important distinction is
between policy output and policy outcome. In the former we are interested in
comparing political measures as such, whereas in the latter we are interested
in comparing their effects, such as the effects of a certain policy measure on
unemployment levels. Outcome data are often easily available, usually in
metrical form and often in time series, and are directly comparable. Compar-
ing output data over time and countries, however, poses more problems.
Therefore, I will concentrate on policy output data here, in order to show the
additional complications in comparing similarity. The methodological
problems discussed later on, however, arise with both output and outcome
data.
A policy field, such as industrial policy, agricultural policy, internal
security policy or environmental policy, consists of a large number of policy
measures. Assessing the similarity of a policy field implies a comparison of
the composition of the whole field: Does the same policy field in two countries
consist of a similar number and type of policy measures? Assessing the simi-
larity of single policy measures implies, for example, asking about the exist-
ence of such a measure, the exact goals pursued with it, or the instruments
used. Thus, to assess policy similarity, one can compare not only the policy
repertoire of a country but also the presence of certain policies, the instruments
used, and sometimes the exact metrical setting of the policy (such as a tax or
European Union Politics 7(2)
272
To continue reading
Request your trial