PLEASE SCROLL DOWN FOR ARTICLE
This article was downloaded by: [Swets Content Distribution]
On: 24 April 2009
Access details: Access Details: [subscription number 902276281]
Publisher Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK
Democratization
Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t713634863
The three worlds of post-communism: revisiting deep and proximate
explanations
Jørgen Møller a
; Svend-Erik Skaaning a
a
Assistant Professor at the Department of Political Science, Aarhus University, Århus C, Denmark
Online Publication Date: 01 April 2009
To cite this Article Møller, Jørgen and Skaaning, Svend-Erik(2009)'The three worlds of post-communism: revisiting deep and proximate
explanations',Democratization,16:2,298 — 322
To link to this Article: DOI: 10.1080/13510340902732565
URL: http://dx.doi.org/10.1080/13510340902732565
Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf
This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.
The three worlds of post-communism: revisiting deep and
proximate explanations
Jørgen MølleraÃ
and Svend-Erik Skaaningb
a
Assistant Professor at the Department of Political Science, Aarhus University,
Bartholins Alle´, Bygning 1331, 8000 A˚rhus C, Denmark; b
Assistant Professor at the
Department of Political Science, Aarhus University, Bartholins Alle´, Bygning 1331, 8000
A˚rhus C, Denmark
(Received March 2008; final version received September 2008)
Since the upheavals of 1989–1991, the post-communist countries have
embarked upon three distinct political trajectories: a path leading to
democracy in the Western part of the setting, a path leading to autocracy in
the Eastern part of the setting, and an intermediate path – both in
geographical and political terms – leading to ‘defective’ democracy. This
article seeks to explain the emergence of these three worlds of postcommunism.
Using typological theory as the principal methodological tool,
we revisit Herbert Kitschelt’s distinction between deep (structural) and
proximate (actor-centred) explanations. The empirical results show that the
post-communist setting is characterized by striking regularities in the form
of clustering in the explanandum as well as the explanans. The orderings of
referents on both the deep and the proximate attributes show a remarkable
co-variation with the political pathways of post-communism – and with
each other. The presence of such systematic empirical regularities lends
support to two conclusions. First, both kinds of explanations elucidate
the present variation in post-communist political regime types. Second, the
variation on the deep factors largely explains the variation on the proximate
factors. Kitschelt’s general plea to dig deeper is thus supported, and the
explanatory quest turns into a challenge of theoretical integration.
Keywords: post-communism; democracy; deep and proximate explanations;
causality; configurational methods
The breakdown of communism in 1989–1991 took an entire generation of scholars
by surprise.1
A fortiori, the nascent field of ‘post-communist studies’ was bereft
of any dominant explanatory paradigm. Not surprisingly, then, the early 1990s
witnessed a plethora of competing theoretical diagnoses of the new reality –
both regarding the direction of change and with the respect to the drivers
of change. As Valerie Bunce has described, optimists pointed to the coming of
ISSN 1351-0347 print/ISSN 1743-890X online
# 2009 Taylor & Francis
DOI: 10.1080/13510340902732565
http://www.informaworld.com
Ã
Corresponding author. Email: jm@ps.au.dk
Democratization
Vol. 16, No. 2, April 2009, 298–322
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
a glorious democratic future based on liberal principles; pessimists foresaw an
unenviable era of political and economic populism, if not outright dictatorship.2
The empirical reality overturned both sets of predictions, however. First,
the setting came to exhibit a fairly systematic partition between democracies
and autocracies; the former cluster situated on the western fringes of the old
empire, the latter cluster inhabiting the eastern territories. Second, a number
of post-communist countries – conveniently situated between the two others
geographically – drifted toward a third cluster of ‘defective democracies’.3
The present article sets out to discuss and elucidate the advent of this tripartite
division of the explanandum, the three worlds of post-communism from which
the title is drawn.
That objective brings us back to the early debate concerning the explanans.
Intertwined with the hefty disagreement about the future shape of the former
Eastern bloc, advocates of very diverse approaches were engaged in rampant
explanatory debates in the early 1990s. At the end of the decade, however, two
competing paradigms had crystallized.
On the one hand, proximate explanations – that is, actor-centred approaches –
were going from strength to strength. The causal factors emphasized by these
approaches were, inter alia, the outcome of the initial elections, constitutional
engineering, and the character of the economic reform process.4
In general, the
matching causal chain led from actor-choices to political outcomes. More particularly,
the proximate explanations of regime change shared two propositions
concerning social change – and in doing so they were heavily indebted to the
so-called school of Transitology which had emerged in the 1980s.5
First, the
important factors shaping the political outcomes dated from the transitional
upheavals, not from antecedent structural factors. Second, and consequently,
these constraints were put in place by actors in a relatively voluntaristic way.6
On the other hand, spearheaded by Herbert Kitschelt’s forceful critique of
these very theories,7
another cohort of students of post-communism were
turning their attention to the way historical legacies and other structural factors
have shaped the scope of choice.8
Kitschelt’s own appraisal is based on a distinction
between deep, structural explanations on the one hand and shallow, proximate
explanations on the other. He anchors the separation on the temporal dimension,
stressing that the deeper the causes are, the more distant they are from the
explanandum, and the more blurred the affiliated causal mechanisms tend to be.9
Contrariwise, proximate causes are situated relatively close to the outcome temporarily
and provide clear-cut causal mechanisms. At first sight, they thus have a
competitive edge as explanans. But looks may be deceiving. Kitschelt’s ontological
point is that the proximate explanations are often too closely (sometimes almost
tautologically) linked with the outcome to be causally interesting vis-a`-vis their
deeper counterparts. As causes, they are better construed as intermediate links in
a chain that leads from the deep factors to the outcome. And this equals saying
that the major theoretical and empirical task is not one of matching the relative
explanatory power of competing bids, but one of theoretical integration.
Democratization 299
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
Somewhat surprisingly, in the context of the former communist countries there
is a lack of research dedicated to assessing the empirical relevance of Kitschelt’s
fundamental critique. In fact, his propositions have not really been tested by
juxtaposing deep and proximate causes in a common explanatory framework,
thereby elucidating their mutual relationship as well as their individual political
effects. The objective of this article is to provide such a test. Our preliminary
assertions, inspired by Kitschelt, can be formulated as a two-sided hypothesis,
amenable to empirical testing:
1) Both the deep and the proximate explanatory variables account for the
present variation in post-communist political regimes.
2) The deep explanatory variables account for the variation in the proximate
explanatory variables.
Methodological issues
That Kitschelt’s critique has not been assessed more systematically probably owes
much to the methodological issues implied by his ontological propositions. As
described above, his recommendations are based on the premise that proximate
explanations are too superficial, temporally speaking. Yet there is more to it than
that. Pari passu, he makes the case for a path-dependent account of political
change in the former communist bloc. Structural conditions to a large extent determined
whether ‘open’ or ‘closed’ politics characterized the transition, the argument
goes, and these characteristics then paved the way for a number of proximate
mechanisms of transmission. From a democratic perspective the post-communist
reality has therefore been one of either ‘virtuous’ or ‘vicious’ circles; a set of positive
and negative spirals unleashed by the deep constraints and then reproduced by
the proximate choices. To elaborate, an auspicious structural point of departure has
made for auspicious proximate actor-choices, ultimately locking in an auspicious
political regime form (in casu democracy) – and likewise (or better: contrariwise)
with an inauspicious structural point of departure.
These arguments have salient methodological consequences. In gist, they
undermine – or at least question – the use of what Kitschelt terms a ‘tournament
of variables’, that is, standard, multiple regression techniques. The more proximate
actor-centred variables will tend to wash out the deeper structural variables
merely because they are more closely linked to the outcome on the dimension of
time. But this tells us preciously little about the causal chain tout court.
Also, and equally problematically, the general tendency of the variables to
coincide in the post-communist setting means that we are presented with
massive problems of multi-collinearity.10
In this situation, multiple regression
analysis, at most, allows us to pinpoint the joint importance of packages of
variables, not their individual effects nor their relative importance.
Kitschelt deserves much praise for having brought these issues to the fore.
Yet it seems to us that he does not take the methodological consequence of his
300 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
own line of reasoning. He seems to favour the use of sophisticated statistical
techniques, such as panel regression analysis, to scrutinize post-communist
pathways for the simple reason that they make up what may be termed the
‘industry standard’.11
However, as argued above, the statistical techniques are
not very helpful when it comes to disentangling the causal mish-mash of
post-communism, the helter-skelter of bivariate relationships. If it is possible to
find a comparative method, or a combination of comparative methods, better
suited for testing Kitschelt’s proposition, one should therefore do so.
On the basis of the preceding discussion, three criteria for such methods
must be observed. First, inferential tools that pave the way for identifying
causal pathways by the systematic use of simple logical arguments concerning
temporal developments are direly needed. Second, the striking empirical
regularities of the explanandum, the existence of three separate worlds of
post-communism, point to the presence of manifest co-variations, which call for
explicit methodological treatment. Third, the logic underpinning the notion of
virtuous and vicious circles means that methodological tools capable of handling
conjunctural causation, given a relatively low number of cases, are warranted.
Bearing this in mind, comparative tools placed in the qualitative tradition –
emphasising differences in kind – appear to be more suitable for the task at
hand than standard quantitative techniques. In this article, we employ a multimethod
approach which allows us to appreciate all the said points in the context
of analysing a medium number of cases. First and foremost, we use typological
theory, a qualitative tool that is virtually tailored to identifying causal pathways
whilst keeping the temporal problematique in mind. As George and Bennet
have pointed out,
[a]n important advantage of typological theorizing is that it can move beyond earlier
debates between structural and agent-centered theories by including within a single
typological framework hypotheses on mechanisms leading from agents to structures
and those leading from structures to agents.12
Three typological analyses of post-communist political pathways are used to test
the hypotheses: from a deep perspective, thereby testing the general relevance
of Kitschelt’s structural corrective; from a proximate perspective, in turn testing
whether the actor-centred approaches account for the variation on the dependent
variable; from a joint – or juxtaposed – perspective by construing each of the
two dimensions as ‘packages’ (i.e., composite indices) and contrasting them.
Subsequently, the typological analysis is backed up by a related technique, namely
crisp-set QCA (csQCA), before a statistical path analysis is employed as robustness-test
to check whether the general results are overturned when treating the differences
within the setting as differences of degree. Finally, when analysing our
typological findings, we discuss the logical criteria for establishing causality – and
use these considerations to construct a general model which integrates the deep
and the proximate explanations.
Democratization 301
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
The qualitative part of this multi-methods approach share some affinity with
that proposed in an interesting, recent article by Carsten Schneider and Claudius
Wagemann.13
Schneider and Wagemann also elaborate Kitschelt’s ontological
propositions by investigating deep and proximate explanations separately before
re-integrating them into a common framework through a ‘two-step approach’.
Like Schneider and Wagemann, we have a somewhat broader understanding of
deep explanations than Kitschelt does: we take them to cover all factors relatively
stable and outside the manipulative reach of (current, domestic) actors.14
This is in
contrast to proximate factors, which can fluctuate significantly within short periods
of time and are easier to change by human agency besides being causally closer to
a particular outcome.
However, we part way with Schneider and Wagemann in a number of
important regards. First, concerning the explanandum, we seek to account for
the kind of regime form, not the extent of democratic consolidation, which is
also reflected by our trichotomizing the dependent variable. Second, whereas
Schneider and Wagemann analyse a more general universe of transitional
countries,15
we are only preoccupied with the post-communist microcosm, the
one Kitschelt had in mind when making his distinction between deep and
proximate causes. Third, and most importantly, Schneider and Wagemann’s
objective is basically to develop the fuzzy-set QCA (fsQCA) approach, using
democratic consolidation merely as an illustration. Ours is to clear empirical
ground – overgrown with a distinct set of causes – using typological theory
backed up by csQCA and path analysis.
The three worlds of post-communism
In spite of what seemed a common point of departure, viz., a patchwork of
interwoven political, social, and economic characteristics,16
the three worlds
of post-communism came into their own very quickly.17
To appreciate this
shift from communist uniformity to post-communist diversity, we set out to
capture the present dividing lines between different post-communist political
regime forms. Doing so, we follow Robert A. Dahl’s procedural definition of
democracy (or polyarchy) accentuating a list of necessary, institutional properties:
elected officials; free, fair, and frequent elections; freedom of expression;
alternative sources of information; associational autonomy; and inclusive
citizenship.18
To measure the status of the empirical referents on these properties, we turn to
Freedom House’s disaggregated data, reported in the Freedom in the World
survey. More specifically, we use the scores of four attributes matching Dahl’s
criteria, viz., electoral process, political pluralism and participation, freedom of
expression and belief, and associational and organizational rights. All of them
are, in accordance with the Dahlian point of departure, considered to be mutually
constitutive attributes. It is therefore proper to construct a composite measure
using a minimum aggregation procedure.19
302 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
After a standardization of the scores (0–100), an application of this logic
produces the following classificatory schema. We consider the countries in the
upper quartile (above 75) to be democracies, the countries in the two lowest
quartiles (below 50) to be autocracies, and the countries in between to be defective
democracies. Since we are interested in the general political demarcation lines of
the setting – rather than finely graded differences – this tripartite classification
should do.20
In the first year for which the disaggregated Freedom House scores
are available, the 2006-scores covering the year 2005, the consequent empirical
ordering is the following.21
Table 1 shows that in 2005, 11 post-communist countries were democracies,
nine were autocracies, and six were defective democracies. Notice, thus, that
although the speed and timing of democratization in individual post-communist
countries have varied, it now appears to have produced distinct forms of political
regimes, coinciding with three geographical subregions. To elaborate, all democracies
are situated in the Western part of the setting, all the autocracies are situated
in the Eastern part of the setting, and – Mongolia being the only exception – the
group of defective democracies in an intermediate area in the south-east. The
descriptive analysis therefore supports the presence of a tripartite empirical
division of the explanandum: a classificatory pattern that has bound in space
and has become more clear-cut over time.
Identifying explanatory variables
As should be clear from the preliminary theoretical and methodological discussion,
the aim of this article is not to find the ‘smoking gun’ – the explanatory variable
– of post-communist democratizations. In fact, we have argued that chasing
such monocausality is not worth while due to the existence of virtuous and vicious
circles. The aim is much more general, viz., to juxtapose and possibly integrate
deep and proximate explanations within the literature, thereby testing the merits
of Kitschelt’s influential critique. Notice that the consequent analysis is very
much theory-driven. Beyond the theoretical objective per se, this is also important
methodologically.
The use of typological theory is premised upon constructing categorical
classifications on the explanatory as well as the dependent variables. When
Table 1. The post-communist distribution of political regime types, 2005.
Democracies Defective Democracies Autocracies
Bulgaria, Croatia,
Czech Republic, Estonia,
Hungary, Latvia,
Lithuania, Poland, Romania,
Slovakia, Slovenia
Albania, Georgia,
Macedonia,
Moldova, Mongolia,
Ukraine
Armenia, Azerbaijan,
Belarus, Kazakhstan,
Kyrgyzstan, Russia,
Tajikistan, Turkmenistan,
Uzbekistan
Democratization 303
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
using such simple dichotomies and trichotomies, in which the cut-off points
become critical whenever the variables are not naturally nominal-scaled, it is
helpful to embed the taxonomic orderings in theory. Consequently, where at all
possible, we will use the data and the cut-off points of existing explanations.
This is also reflected by the fact that we dichotomize ordinal- and intervalscaled
variables using thresholds justified by the extant theories, not around
means or medians.
More to the point, to facilitate the subsequent typological analysis, all the
variables are dichotomized in such a way that the presence of a particular attribute
implies the theoretical expectance of democracy whereas the absence of the property
implies the expectance of autocracy. A thorough reading of the literature
on post-communism22
has identified four deep and three proximate explanatory
variables, which are presented and operationalized below.
Political legacies (LEGACIES) as described by Herbert Kitschelt.23
Based
on the status on the two pre-communist attributes of bureaucratic state legacies
and the balance of power between communists and their challengers at the
introduction of communist rule, Kitschelt makes a distinction between the
respective legacies of i) bureaucratic-authoritarian, ii) national-accommodative,
iii) patrimonial communism, and iv) colonial periphery. The argument holds
that both bureaucratic-authoritarian and national-accommodative communism
made for a democratic transition in 1989–1991 due to the strength of the
opposition. In contrast, patrimonial communism and its sub-class, colonial
periphery, made for autocratic continuity because the communist incumbents
only faced weak opposition. We dichotomise the variable accordingly.
Modernization (MODERN) as used by Marcus J. Kurtz and Andrew Barnes.24
According to modernization theory, socio-economic development promotes
democracy through the empowerment of civil society, thus increasing tolerance,
moderation, and the demand for freedom. We dichotomize the variable on the
basis of the first reported GDP figures from the World Bank (PPP, constant
2000 international USD) after the breakdown of communism – and the dissolution
of the Soviet Union and Yugoslavia – using a threshold of $5300 per capita. This
threshold is contained within what may be termed a zone of relative affluence and
is well-suited as cut-off point empirically.25
Vicinity to Western Europe (WEST) as described by Jeffrey S. Kopstein and
David A. Reilly.26
Kopstein and Reilly use the distance of a post-communist country’s
capital to the twin capitals of Berlin and Vienna – whichever is closer – to
measure this variable. They also operate with a distinct neighbour-effects variable
based on the average democracy-score in neighbouring countries. It turns out that
the two correlate overwhelmingly (Pearson’s r¼0.93). To avoid awarding double
weight to the geographical factor, we have therefore confined our attention to the
former attribute. It has been dichotomized using a four-fold distinction made by
Kopstein and Reilly in an overview table, in which they draw borderlines using
the three distances of 500, 1000, and 1500 miles. We collapse the two first of
these categories, expected to be positively associated with democracy, and the
304 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
two last, expected to be negatively associated with democracy. Hence, 1000 miles,
placed in a wide, natural gap in the empirical distribution, constitutes the cut-off
point.
The Resource Curse (NOOIL) as described by Theresa Sabonis-Helf and
Younkyoo Kim.27
The logic of this explanation is that governments are naturally
induced to use their revenues from natural resources, first and foremost oil, to
undermine democratization. The money comes with no strings attached and can
therefore be used to relieve social pressures through the provision of patronage
and public goods – and at the same time it opens a wide avenue for repressing
oppositional forces. Fine-grained data on oil production are unfortunately rather
unreliable.28
This does not present much of a problem here, however, as heavy
public reliance on oil money is very much a matter of ‘either-or’, not least in
the context of the examined countries. We order the countries dichotomously by
isolating the post-communist hydro-carbon rich countries listed in the IMF
Guide on Resource Revenue Transparency.29
Displacement of communist incumbents at the first elections (DISPLACEMENT)
as described by M. Steven Fish.30
Fish argues that the outcome and
the character of the initial elections after the breakdown of communism laid
out the tracks of the economic reform process. Others have extended the causal
chain to the political reform process, arguing that an oppositional win over the
communist incumbents at the first elections favoured democratization.31
Generally
speaking, Fish’s original data is well-suited for our purposes. However, to
avoid tautological reasoning arising from the fact that some of the subcomponents
of his composite index of displacement can be construed as
proxies of democracy, we will only use the sub-component measuring the
actual outcome of the initial elections. This ordering is easily converted into a
dichotomous classification as Fish’s score of 0 indicates the absence of incumbent
displacement, whereas the two other possibilities (1 and 2) indicate some sort of
displacement.
Economic reform (REFORM) as described by M. Steven Fish and Omar
Choudhry.32
Entering the debate pitting economic shock therapy against gradualism,
Fish and Choudhry argue that economic liberalization has had a positive
effect on democratization in the longer term. The numbers measuring economic
liberalization are derived from De Melo et al., who – in turn – distinguish
between advanced reformers, high-intermediate reformers, low-intermediate
reformers, and slow reformers.33
We use this distinction between reformers (the
former two classes) and non-reformers (the latter two classes) to dichotomize
the variable.
Strong Legislatures (LEGISLATURE) as described by M. Steven Fish.34
Fish
creates a Parliamentary Power Index, on a scale from 0 to 1, with higher scores
indicating stronger legislatures. He then argues that a high degree of parliamentary
power (and hence a weak presidency) makes for democracy whereas a low degree
of parliamentary power (and hence a strong presidency) makes for autocracy.
More particularly, Fish emphasises that a country that opted for a strong legislature
Democratization 305
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
is one that scored above 0.60 in the Index.35
We dichotomize the variable using
this threshold.36
Testing the deep explanations
To test the merits of the deep and proximate explanations, we turn to typological
theory. A brief note on this technique seems warranted. A typology is a multidimensional
and conceptual classification, that is, it is an ordering on a compound
of attributes.37
Yet the pure ordering of a multi-dimensional property space is
merely the descriptive face of typologies. Typologies also offer an anchorage for
causal inference.38
This is so because the respective orderings on attributes, of
which the compound is made up, normally capture a theoretically relevant
distinction between the presence and absence of an explanatory factor or the
phenomenon to be explained. A typology formed in this way therefore necessarily
posits something about theoretical relationships within the property space at
hand – it delineates the expected pathways within the typology itself.39
On the basis of the serial operations on the four structural variables and the
dependent variable, it is straightforward to form the deep typology by unfolding
the complete property space of 48 types. However, before doing so, the
theoretical expectations must be spelled out. Recall that the justifications for
including each of the explanatory variables was that when the attribute used
to separate the dichotomous classes was present, the variable would theoretically
make for democracy and when absent it would theoretically make for
autocracy.
The following empirical expectations can be formulated. First, democracy and
autocracy are thought to reflect two equilibrium-points in which all deep attributes
are present and absent, respectively. By implication, and bearing the
relatively large number of democracies and autocracies in mind, the prediction
is that many of the cases clump in the two polar types of cell 1 and 48.40
We
refer to the former polar type as ‘structure-based democracy’ and the latter
polar type as ‘structure-based autocracy’. Second, the defective democracies
should inhabit the intermediate types close to the diagonal reflecting a mix of
present and absent attributes. In Figure 1, the typology is depicted with empirical
referents.41
By and large, the ordering matches the theoretical expectation, mirroring the
fact that the countries clump along the diagonal. Nine countries belong in the
polar type of structure-based democracy and four countries in the polar type of
structure-based autocracy. Moreover, none of the six defective democracies
either has or lacks all the structural attributes. Consequently, 19 out of the 26
countries are classified as anticipated theoretically. Notice, in addition, that most
of the Western countries are to be found in polar type 1, most of the Eastern
countries in polar type 48 or in adjacent types, and many of the ‘midway’ countries
of the south-east in the interposed types – reflecting the three worlds of postcommunism.
Viewed from the higher ground, then, the analysis confirms the
306 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
relevance of understanding the post-communist political pathways through the
prism of a deep analysis.
Testing the proximate explanations
On the basis of the dichotomous classifications of the three proximate variables
and the trichotomy of political regime forms, a typology consisting of 24 types
can be unfolded. The theoretical predictions mirror those of the preceding analysis.
Figure 1. The full deep typology with empirical referents, 2005.
Democratization 307
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
Where all three attributes are present, democracy is the expected political
outcome, and where all three attributes are absent we anticipate an outcome of
autocracy. Any other combination implies an expectation of defective democracy.
This time we name polar type 1 ‘actor-induced democracy’ and polar type 24
‘actor-induced autocracy’. Figure 2 shows to what extent the 2005-reality of
post-communism conforms to the predictions.
Nine countries belong in the type of actor-induced democracy and six
countries in the type of actor-induced autocracy, that is, in the two combinations
construed as stable equilibria. Of the six defective democracies, four have the
expected mixed combinations of attributes whereas two (Macedonia and
Ukraine) exhibit the respective presence and absence of all attributes, showing
that the fit is not perfect.
Still, when considering the theoretical expectations, no less than 19 out of 26
countries are again classified as expected. This is also visibly to the naked eye as
Figure 2. The full proximate typology with empirical referents, 2005.
308 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
the countries clearly clump along the diagonal. Notice, once more, that most of the
Western countries are placed in polar type 1 whereas most of the Eastern countries
group in polar type 24. Finally, a large proportion of the ‘intermediate’ countries
are found in the interposed types, yet again highlighting the existence of three
worlds of post-communism.
Juxtaposing and integrating the two explanatory packages
A typological analysis
The two preceding analyses both exposed some striking empirical regularities. Not
only was it possible to draw a very coherent picture of the causal pathways, it was
possible to do so from a deep as well as a proximate point of view. Fact of the matter
is that the explanatory power of the two approaches appears even-handed – 19 out
of 26 referents were classified in accordance with the predictions both times
around. The only substantial difference between the approaches was that relatively
more defective democracies were ‘correctly’ classified by the deep typology and
relatively more autocracies by the proximate typology. Taken together, the analyses
therefore supported the first of our two hypotheses – that both the deep and the
proximate explanations explain the present variation in post-communist political
regime types. Harking back to the second hypothesis, the interesting question is
now: How do the deep and proximate variables, taken as wholes, relate to each
other?
To juxtapose the deep and proximate variables, it makes sense to treat each
of the two clusters of variables as packages (i.e., as composite indices). The
justification for doing so is the striking regularities discovered in the preceding
sections. To capture these regularities, we construe each of the clusters of
variables as one dimension on which we may either encounter the presence of
all attributes, the presence of some but not all attributes, or the absence of all
attributes.42
Theoretically, a full presence connotes that the country in question
is expected to be a democracy, a mixed score that it is expected to be a defective
democracy, and a full absence that it is expected to be an autocracy.
When combining these two trichotomous classifications, a typology once
again comes into existence. This final typology exhausts the possible logical combinations
of statuses on the two explanatory dimensions. In Figure 3, the typology
is illustrated with empirical referents.
One thing is immediately clear. The orderings on the two dimensions correlate
very strongly.43
Thirteen out of 26 countries are found within the two polar types.
An additional nine falls into the mixed-mixed type, which means that only four
do not clump along the diagonal of the typology. Notice, furthermore, that all
the deviant cases fall into types adjacent to the types making up the diagonal.
No country thus falls into the theoretically implausible types that combine a full
presence on one package with a full absence on the other package.
We also see clear evidence of the three worlds of post-communism. In
Figure 3, bold is used to highlight the democracies and italic to highlight the
Democratization 309
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
autocracies. The clear majority of the democracies fall into polar type 1, which we
term ‘guaranteed democracy’. Likewise, four out of nine autocracies are found in
polar type 9, which we term ‘guaranteed autocracy’. Finally, four out of six defective
democracies are contained in type 5, the intermediate type par excellence.
Aberrations are encountered, however, as a good handful of countries have
become either democracies or autocracies in spite of not having or not lacking
all attributes. We will discuss some of these in the context of the subsequent
csQCA-analysis. But at this stage a very simple observation, linked to the classificatory
logic, is pertinent. Recall that classes are separated by differences in kind
whereas the differences within a class are measured in degree. Consequently, we
Figure 3. The full typology of deep and the proximate packages with empirical referents.
310 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
can elucidate the differences in the degree of democraticness and autocraticness
within the twin classes of democracy and autocracy by scrutinizing the
disaggregated Freedom House ratings. One thing is noteworthy: The democracies
that lack some attributes on both dimensions (Bulgaria and Romania) generally
score lower on our democracy measure than the democracies with full presence
of attributes. Likewise, the autocracies sharing some attributes on both dimensions
(Armenia, Kyrgyzstan, and Russia) generally score higher than the autocracies
with full absence of attributes – or merely a full absence on one of the dimensions.
Accordingly, these few outliers are, each from their side, closer to the thresholds
demarcating the class of defective democracy than their political cell-mates, just
as one would expect.
The most important task here, however, is to relate the two alternative kinds
of approaches to each other in order to integrate them into a common explanatory
framework. Recall in this connection that causality is a concept implying at
least four requirements: 1) a theoretical link between cause and effect; 2) a corresponding
empirical co-variation; 3) that the cause antedates the effect; 4) that
alternative explanations are controlled for.44
Both the deep and the proximate
variables fulfil each of these criteria (as causes). The respective variables were,
first, selected with reference to their theoretical relationship with the outcome in
question, second, turned out to stand in the expected empirical relationship to
the three possible outcomes on this variable, third, were measured at a point in
time predating the measurement on the dependent variable, and, fourth, were
chosen from the plethora of theoretical propositions based on their explanatory
power in relation to their alternatives in previous studies.
Yet much the same can be said about the relationship between the deep variables
and their proximate counterparts. The structural attributes are all more basic
(deep-seated) than the actor-centred. Moreover, the two packages stand in a very
systematic empirical relationship and it hard to imagine any alternative factor that
could spoil this. Can we, finally, establish a theoretical link between the two
packages? Clearly we cannot create a causal chain leading from the choices of
the actors to the structural attributes because of the temporal sequence. Contrariwise,
it seems very much possible to establish a causal chain leading from the deep
attributes to the proximate actor-choices.
Features such as the outcome of initial elections, constitutional choice, and
economic reforms to a large degree capture various aspects of ‘open politics’
and societal mobilization. It is very likely that the structural point of departure
should impact on this. If a society is modernized, has a vibrant civil society as
well as cultural and political linkages to Western Europe, and is not subjected to
the resource curse – as was (and is) the case in the Western part of the
post-communist world – competitive politics, including the formation of viable
oppositional parties, is to be expected during a transitional window of opportunity.
If, on the other hand, all of these (or almost all of these) attributes are absent – as
was (and is) the case in the Eastern part of the post-communist world – competitive
politics is not to be expected in such a situation.
Democratization 311
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
In a nutshell, the actors do not act in a vacuum, and the ‘deep arena’ is likely to
significantly shape their room for action. Even though the causal chain may very
well work through the proximate variables, they should therefore not be regarded
as root causes, but rather as mechanisms of transmission. In sum, then, we posit
that the deep structural point of departure has determined, or at least very much
constrained, the proximate choices of the actors and that the combination of
these two dynamics have shaped the political pathways of post-communism – a
cross-temporal dynamic that the notion of virtuous and vicious circles conjures
up an apt image of.
One should not get carried away by the merits of this explanatory edifice,
though. As Kitschelt has pointed out, the actor-centred choices may be completely
spurious. To quote:
Either the deeper causes x ‘work through’ the shallower cause y to bring about the
final outcome z (x!y!z); or the cause x bring about both what appears as the
shallower cause y as well as the outcome z (x!y; x!z).45
This objection goes for our analysis, too. As a result of the high collinearity,
viewed from the higher ground it is very difficult to distinguish between these
alternatives. The ultimate test requires actual process-tracing of the individual
post-communist cases,46
but such an exercise lies beyond the scope of this
article. Notice, however, that the convincing theoretical links between the identified
proximate attributes and the outcome on the dependent variable supports
the notion that some of the causal impact does indeed pass through these
choices. Furthermore, the feasible theoretical links between the deep structural
constraints and the proximate actor-choices further underpin the notion of a
coherent chain – of an integrated whole that is.
Are the results robust?
In order to re-test the results derived from the typological analysis, we first employ
csQCA to scrutinize if the same empirical patterns emerge from an alternative,
although associated, kind of data treatment.47
CsQCA is a configurational
method which – in the context of the social sciences – represents an attempt to
narrow the gap between quantitative (variable-oriented) and qualitative (caseoriented)
research.48
It is based on set theoretical reasoning (Boolean algebra)
and, consequently, allows the researcher to identify multiple conjunctural relationships
in terms of necessity and/or sufficiency. Like case-oriented approaches, it
treats cases as wholes, that is, the different aspects of a case are defined in relation
to each other. At the same time, however, it shares the variable-oriented techniques’
broad understanding of social phenomena – and embraces their ability
to reduce complexity through mathematical data ‘manipulation’. For readers not
familiar with crisp-set analysis, when reading the formalized results, upper-case
letters indicate the presence of an outcome or causal condition (coded 1) and
312 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
lower-case letters indicate its absence (coded 0). Concerning the interpretation of
Boolean operators, Ã
means logical and,þmeans logical or.
As the required dichotomous coding of the outcome (democracy) and the conditions
has already been carried out to facilitate the typological analysis, we can
move straight to the construction of a truth table. This is achieved by regrouping
all identical cases, in terms of their scores on the conditions, into a single configuration.
The consequent truth table (Table 2) shows that the 26 cases are covered
by 12 different configurations.
No contradictions are uncovered, that is, no cases have identical conditions but
different outcomes. We are therefore able to further reduce complexity through the
application of the so-called minimization rule. The rule says that if two Boolean
expressions (configurations) differ in only one causal condition but produce the
same outcome, the causal condition that distinguishes the two expressions is
irrelevant and can be removed to create simpler, combined expressions.49
The truth table shows that no less than five conditions, viz., vicinity to Western
Europe, modernization, no large-scale oil production, a strong legislature, and
extensive economic reforms are present in both of the configurations linked to a
positive outcome. They can thus be identified as necessary for democracy in the
post-communist context. However, in order to constitute a sufficient path (to
democracy), they have to be combined with either the simultaneous presence
of favourable political legacies and displacement of the incumbents in the first
election or, alternatively and somewhat counter intuitively, the absence of these
Table 2. Truth table of the dichotomized data linked to outcome and conditions.
Country Demo. Legac. West. Mod. Nooil Legis. Displ. Ref.
Croatia, Czech
Republic, Estonia,
Hungary, Latvia,
Lithuania, Poland,
Slovakia, Slovenia
1 1 1 1 1 1 1 1
Bulgaria, Romania 1 0 1 1 1 1 0 1
Macedonia 0 0 1 1 1 1 1 1
Albania 0 0 1 0 1 1 0 1
Moldova 0 0 1 0 1 1 1 0
Mongolia 0 0 0 0 1 1 0 1
Ukraine 0 0 1 1 1 0 0 0
Belarus 0 0 1 0 1 0 0 0
Russia 0 0 0 1 0 0 1 0
Armenia, Georgia,
Kyrgyzstan
0 0 0 0 1 0 1 0
Tajikistan 0 0 0 0 1 0 0 0
Azerbaijan,
Kazakhstan,
Turkmenistan,
Uzbekistan
0 0 0 0 0 0 0 0
Democratization 313
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
two attributes. In formal terms, the paths explicitly connected with democracy
look like this:50
1) WESTÃ
MODERNÃ
NOOILÃ
LEGISLATUREÃ
REFORMÃ
LEGACIESÃ
DISPLACEMENT (Croatia, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland,
Slovakia, Slovenia) (82%) þ
2) WESTÃ
MODERNÃ
NOOILÃ
LEGISLATUREÃ
REFORMÃ
legaciesÃ
displacement (Bulgaria, Romania) (18%).
In toto, these results51
lend strong support to our conclusions based on the typological
analyses. First, the full model does not produce many contradictions; in
fact it produces none – which mirrors the presence of manifest regularities on
the ground. Second, all conditions are included in the paths, showing the expected
direction, which indicates that they play a significant role in the virtuous circles
described in this article. Third, both deep and proximate conditions are identified
as critical, underscoring that conjunctural causality rather than monocausality is a
fact of life in the post-communist world.
The csQCA results further emphasise an interesting point which was also
evident in the typological analyses – that Bulgaria and Romania are ‘democratic
overachievers’ considering their statuses on the explanatory attributes. To be sure,
as we have already noted they are less democratic than their cell-mates within
the class of democracy. Nevertheless, both countries have been able to sustain
democracy in spite of their lack of a full presence of attributes on both the deep
and the proximate package. The surprising merits of Bulgaria and Romania can
probably – partially at least – be explained by a factor we have not considered
as it did not fit into the distinction between deep and proximate causes,52
viz.,
the EU-enlargement process.
As Milada Anna Vachudova53
has convincingly argued, Bulgaria and
Romania only made the democratic turn after the ascendancy of what she terms
the ‘active’ leverage of the EU, i.e., after the initiation of actual membership
negotiations in the mid-1990s. In the early 1990s, by contrast, the lack of political
competition allowed the communist incumbents in these countries to pursue an
illiberal course of action; something the then ‘passive’ leverage of the EU could
not hinder. Translated into our explanatory edifice, the deep constraints – in particular
the political legacies – were unfavourable to democracy. Only the impact
of the enlargement process game made it possible to break these constraints. This
goes to show that more attention could be paid to the relative weight of various
explanatory factors over time, at least when descending the ladder of abstraction
and scrutinizing particular countries.
As the last step in a multi-methods approach, we employ a statistical path
analysis to further control the reliability of the typological results. First, a latent
structural dimension and a latent actor-centred dimension are extracted from the
raw data using principal component analysis. One factor with an eigenvalue
above one is extracted for each group of variables; for both of them it accounts
314 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
for approximately 70% of the variance. The factor scores derived from a principal
component analysis (regression method) is then brought into play in a path
analysis. Path analysis makes it possible to identify causal linkages among a set
of variables. It is thus suitable for testing a causal model from a given theoretical
stance – i.e., with the previous research on the area in mind – as it provides an
indication of the model’s strength.
The results are illustrated in Figure 4. It turns out that the deep package has a
direct impact on democraticness as well as and an indirect impact through the
proximate package. The path coefficients reflect the amount of variability in one
factor that is explained by the other. In brief, the model indicates that the structural
package predicts 78% of the variation in the actor-based package, and that the
direct explanatory power of the two packages amounts to 22% and 71% of the
variation in democraticness, respectively.54
The high (adjusted) R-square of
0.78 tells us that the overall model provides a very good account for the variability
in post-communist democraticness.
It is evident that the path analysis does not contradict our previous findings. On
the contrary, the statistical model reconfirms the principal results of the typological
analyses as both packages show a strong relationship with the extent of democraticness
and, moreover, as the structural package explains most of the variation in
the actor-centred package. The prior results are therefore not overturned – indeed,
they are clearly supported – when construing the differences within the setting as
differences of degree only. Finally, the model is a fine illustration of one of our
initial methodological points. It shows that a classical ‘tournament of variables’
approach would have obscured the identified causal chain as the proximate
package more or less wash away the deep package, statistically speaking.
Conclusions
This article set out to explain the advent of three separate worlds of postcommunism
– a systematic geographic distribution of political regimes – less
than two decades after a seemingly uniform point of departure. Building on
Kitschelt’s influential paper from 2003, we have demonstrated that the most
salient fact about the post-communist political pathways is the striking regularities
we encounter, both from and a deep and a proximate point of view. But this means
Figure 4. Path diagram illustrating the link between the explanatory packages and
democraticness.
Democratization 315
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
that we are left with a Gordian knot as we have to consider which set of factors we
should pin our faith in.
How to cut the knot? Our solution is fairly simple as we reject wielding the
sword at all. Instead of opting for one package only – or one explanatory
factor – the proximate attributes should be construed as intervening links in a
causal chain that leads from the deep attributes to the political outcome. Structures
do not create democracy (or autocracy for that matter); actors do. But the
systematic – as opposed to random – diversity of post-communism, extending
to the explanandum as well as the explanans, can only be thoroughly explained
with reference to deeper factors. In a nutshell, our argument is that these are the
genuine explanatory variables but that they kick in via the causal mechanisms
provided by the proximate explanations.
As such, the empirical analyses have confirmed the two-sided hypothesis formulated
in the introduction. The theoretical claim that arises from this conclusion
is that virtuous and vicious circles, respectively, characterize the relationship
between the deep and proximate factors, thereby conforming to a path-dependent
logic. This equals saying that the dividing lines between democracies (in the
Western part of the setting) and autocracies (in the Eastern part of the setting)
are if not set in stone then at least very clear-cut. Methodologically, the contrast
between these equilibrium-points, mirroring the respective presence and absence
of the theoretically interesting attributes, are therefore best described by differences
in kind, rather than differences in degree; which means that classificatory
schema and typological mappings are appropriate when seeking to navigate the
post-communist waters.
One important qualifications needs to be made here, however. The analyses
also indicated that the class of defective democracy is characterised by diversity
rather than uniformity. Although many of the countries showing a mixed score
on both the deep and the proximate package were defective democracies, there
were even more exceptions to this rule. Such regimes therefore cannot be construed
as stable equilibria. Rather, they have an intrinsic propensity to create conjunctures
– as has in fact been empirically demonstrated by the ‘colour
revolutions’ in Georgia (2003), Ukraine (2004), and Kyrgyzstan (2005). Why
this is so is a question that will not be addressed here, as it would require a new
article. But one hypothesis that arises from our analysis is the following: the
very fact that these countries have a mixed sheet on the explanatory factors
means that the actors have much more room for manoeuvring than what would
otherwise be the case.55
Even in this case, it is thus possibly to argue that the deep constraints have set
the parameters for the actors. In total, then, the explanatory edifice arrived at
represents a blow against the constricted focus on proximate explanations that
dominated the study of regime change from the mid-1980s till the mid-1990s –
even though it incorporates and relies on the findings of these approaches. For
this research tradition does not construe the actor-centred explanations as intervening
in nature, rather they portray them as the interesting explanatory variables
316 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
while emphasizing the fairly wide room for choice on the actors’ behalf. This
article has demonstrated the limitations of such a perspective, and we conclude
with a plea for greater attention to structural conditions and for a further integration
between deep and proximate theories.
Acknowledgments
A large number of people have assisted us in the work that culminates with this article. This
assistance is acknowledged in our respective doctoral dissertations. Besides that, we are
thankful for the critical comments of two anonymous reviewers. The usual disclaimer
applies.
Notes
1. McFaul, ‘Fourth Wave’, 212.
2. Bunce, ‘Political Economy of Postsocialism’, 758.
3. Cf. Merkel, ‘Embedded and Defective Democracies’.
4. E.g., Fish, ‘Determinants’; Fish, ‘Stronger Legislatures’; Fish, ‘Democratization’s
Requisites’; Fish and Choudhry, ‘Democratization and Economic Liberalization’;
Ishiyama and Velten, ‘Presidential Power’; McFaul, ‘Fourth Wave’; Roeder,
‘Varieties of Post-Soviet Authoritarian Regimes’.
5. See O’Donnell and Schmitter, ‘Tentative Conclusions about Uncertain Democracies’.
6. See, e.g., Fish, ‘Determinants’, 77–8; Fish, ‘Stronger Legislatures’, 11–12.
7. Kitschelt et al., Post-communist Party Systems; Kitschelt, ‘Accounting for Outcomes
of Post-Communist Regime Change’; Kitschelt, ‘Accounting for Postcommunist
Regime Diversity’. Notice that Kitschelt’s 1999 book analyses different patterns of
democratic party competition. In the two subsequent papers, however, his focus is
squarely on explaining democratization, as is ours in this article.
8. A few authors, such as Janos, ‘Continuity and Change in Eastern-Europe’, in fact
anticipated Kitschelt. But many more were to follow in his footsteps, e.g. those
included in Ekiert and Hanson, Capitalism and Democracy in Central and Eastern
Europe; Darden and Grzymala-Busse, ‘Great Divide’; Pop-Eleches, ‘Historical
Legacies’.
9. Kitschelt, ‘Accounting for Postcommunist Regime Diversity’, 74–5.
10. Cf. also Pop-Eleches, ‘Historical Legacies’, 917.
11. This assertion is based on a critical review by Herbert Kitschelt of the doctoral
dissertation of one of the authors in which he recommend – verbatim – for observing
such an ‘industry standard’.
12. George and Bennett, Case Study, 245.
13. Schneider and Wagemann, ‘Reducing Complexity’.
14. Schneider and Wagemann try to capture this by replacing the adjective ‘deep’ with
‘remote’.
15. Including Southern European, Latin American, and some – but far from all –
post-communist countries.
16. Bunce, ‘Paper Curtains and Paper Tigers’, 981.
17. Kitschelt, ‘Accounting for Postcommunist Regime Diversity’, 49.
18. Dahl, On Democracy, 85.
19. According to Gary Goertz, using the minimum score – corresponding to the fuzzy
and operation in fsQCA – with regard to the constitutive features is the preferable
way to handle an aggregation process for many social science concepts. This certainly
applies to the concept of democracy, which he – tellingly – uses to illustrate the
point. See Goertz, Social Science Concepts. For similar arguments, see Bowman,
Democratization 317
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
Lehoucq, and Mahoney, ‘Measuring Political Democracy’, 939–70; Sartori, Theory
of Democracy Revisited.
20. Furthermore, an application of a different dataset concerning the same year, viz., the
Bertelsmann Transformation Index 2006, supports our ordering. Using the minimum
procedure on the Bertelsmann scores connected to the attributes of political
participation (free, fair, and frequent elections, effective power to elected leaders,
freedom of association and assembly, freedom of expression), the exact same distribution
of cases into classes appear – with the sole exception of Macedonia. Based on
this inductive line of reasoning, combined with our case knowledge, the ordering
seems robust and plausible.
21. Due to missing data on the explanatory variables we have been forced to exclude
Bosnia-Herzegovina and Serbia-Montenegro, leaving a set of 26 post-communist
countries to be classified.
22. This reading is based, in particular, on the overviews presented by Kitschelt,
‘Accounting for Postcommunist Regime Diversity’ and Pop-Eleches, ‘Historical
Legacies’.
23. Kitschelt et al., Post-communist Party Systems; Kitschelt, ‘Accounting for Outcomes
of Post-Communist Regime Change’.
24. Kurtz and Barnes, ‘Political Foundations’.
25. As it falls in a natural gap in the empirical distribution of cases.
26. Kopstein and Reilly, ‘Geographic Diffusion’.
27. Sabonis-Helf, ‘Rise of the Post-Soviet Petro-states’; Kim, Resource Curse. See also
Fish, Democracy Derailed and some of the contributions in Ebel and Menon,
Energy and Conflict.
28. Fish, Democracy Derailed, 84.
29. IMF, Guide on Resource Revenue Transparency.
30. Fish, ‘Determinants’.
31. Bunce, ‘Political Economy of Postsocialism’, 782; McFaul, ‘Fourth Wave’; Fish,
‘Democratization’s Requisites’.
32. Fish and Choudhry, ‘Democratization and Economic Liberalization’.
33. De Melo, Denizer, and Gelb, ‘Patterns of Transition’.
34. Fish, ‘Stronger Legislatures’.
35. Ibid., 18.
36. Turkmenistan is not included in Fish’s survey. However, his index is very much of
the same ilk – and on the same scale – as the Presidential Power Index of Frye,
‘A Politics of Institutional Choice’. In Frye’s index, Turkmenistan receives a very
high score of presidential power meaning that it is fair to place Turkmenistan in
the group of countries with weak legislatures. In the statistical path analysis, we
use Frye’s score for Turkmenistan (0.35).
37. Cf. Lazarsfeld and Barton, ‘Qualitative Measurement’; Bailey, Typologies and
Taxonomies.
38. George and Bennett, Case Study; Elman, ‘Explanatory Typologies’.
39. George and Bennett, Case Study, 235.
40. 48 types may strike the reader as quite a complex construct, the danger of which
is often warned about in the literature on typologies (e.g., Elman, ‘Explanatory
Typologies’). However, as we predict a clustering around the diagonal, the presence
of five variables does not present an unwieldy complexity.
41. The respective positioning of the independent variables is coincidental. It could take
many other shapes but this would not change the identity of the 48 types.
42. This equals the typological technique termed ‘indexing’ by Elman, ‘Explanatory
Typologies’.
318 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
43. If we interpret the three values as an ordinal-scale of 1 (full presence), 2 (mixed
score), and 3 (full absence), the statistical correlation (Kendall’s Tau-b) between
the two packages is an impressive 0.86.
44. Cf. Gerring, Social Science Methodology, 130–46.
45. Kitschelt, ‘Accounting for Postcommunist Regime Diversity’, 75.
46. Cf. George and Bennett, Case Study.
47. As this technique is primarily used to check the robustness of the findings, the
description of the different analytical steps and reports of the results are comprised.
We do, however, recognize the standards of good practice – except (due to space
limitations) the publication of the raw data matrix – suggested by Wagemann and
Schneider, Standards of Good Practice.
48. Ragin, Comparative Method; Rihoux and Ragin, Configurational Comparative
Analysis.
49. Ragin, Comparative Method, 93.
50. If we allow all logical cases, i.e., empirically unobserved configurations (combinations
of attributes), in the Booelan reduction (see Ragin and Sonnett, Between Complexity and
Parsimony) this result emerges:
LEGACIES (Croatia, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, Slovenia)þ
MODERNÃ
displacementÃ
LEGISLATURE (Bulgaria, Romania).
As we do not, however, wish to make any conclusions based on assumptions about
unobserved variation, this result is not subjected to further interpretation.
51. CsQCA is only used as a secondary analytical tool. Thus, the space limitations do not
allow us to interpret the results in detail, nor to discuss the complementary paths
leading to non-democracy – apart from mentioning that they definitely do not
contradict our arguments presented in the core text. The absence of favourable
political legacies is the only necessary condition, while the analysis points to no
less than seven different, and rather complex, solution terms:
modernÃ
NOOILÃ
legaciesÃ
displacementÃ
LEGISLATUREÃ
REFORM (AlbaniaþMongolia)þ
modernÃ
NOOILÃ
legaciesÃ
westÃ
legislatureÃ
reform (Armenia, Georgia, KyrgyzstanþTajikistan)þ
modernÃ
legaciesÃ
westÃ
displacementÃ
legislatureÃ
reform (Azerbaijan, Kazakhstan, Turkmenistan,
UzbekistanþTajikistan)þNOOILÃ
legaciesÃ
WESTÃ
displacementÃ
legislatureÃ
reform (Belarusþ
Ukraine)þMODERNÃ
NOOILÃ
legaciesÃ
WESTÃ
DISPLACEMENTÃ
LEGISLATUREÃ
REFORM(Macedonia) þmodernÃ
NOOILÃ
legaciesÃ
WESTÃ
DISPLACEMENTÃ
LEGISLATUREÃ
reform(Moldova)þMODERNÃ
nooilÃ
legaciesÃ
westÃ
DISPLACEMENTÃ
legislatureÃ
reform (Russia).
An inclusion of all logical cases in the minimization procedure produces paths toward
non-democracy has the following consequences:
modern (AlbaniaþArmenia, Georgia, KyrgyzstanþAzerbaijan, Kazakhstan, Turkmenistan, Uzbekistanþ
BelarusþMoldovaþMongoliaþTajikistan)þlegislature (or reform – logically substitute)
(Armenia, Georgia, KyrgyzstanþAzerbaijan, Kazakhstan, Turkmenistan, UzbekistanþBelarusþRussiaþ
TajikistanþUkraine)þ legaciesÃ
DISPLACEMENT (Armenia, Georgia, KyrgyzstanþMacedoniaþ
MoldovaþRussia).
52. Furthermore, the inclusion of the EU-variable into the general explanatory edifice
would imply salient endogeneity problems. To the extent that potential EU membership
is considered to be a structural variable, the variable measuring vicinity to
Western Europe tends to be a fairly good proxy.
53. Vachudova, Europe Undivided.
54. All relationships are statistically significant at the 0.1-level.
55. For an example of a lucid analysis of actor-choices within such structural constraints,
see Hale, ‘Regime Cycles’.
Democratization 319
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
Notes on contributors
Jørgen Møller has a PhD from the European University Institute in Florence, Italy (2007)
and is currently Assistant Professor at the Department of Political Science, Aarhus University,
Denmark. His research interests include conceptualization of democracy, post-communist
political change, patterns of state formation, and qualitative methodology.
Svend-Erik Skaaning has a PhD from the University of Aarhus (2006) and is currently
Assistant Professor at the Department of Political Science, Aarhus University, Denmark.
His research interests include conceptualization of democracy, conceptualization of civil
liberties, political change in developing countries, patterns of state formation and qualitative
methodology.
Bibliography
Bailey, Kenneth D. Typologies and Taxonomies: An Introduction to Classification
Techniques. London: Sage, 1994.
Bertelsmann Transformation Index 2006, http://bti2006.bertelsmann-transformation-
index.de/
Bowman, Kirk, Fabrice Lehoucq, and James Mahoney. ‘Measuring Political Democracy:
Case Expertise, Data Adequacy, and Central America’. Comparative Political
Studies 38, no. 8 (2005): 939–70.
Bunce, Valerie. ‘Paper Curtains and Paper Tigers’. Slavic Review 54, no. 4 (1995): 979–87.
Bunce, Valerie. ‘The Political Economy of Postsocialism’. Slavic Review 58, no. 4 (1999):
756–93.
Dahl, Robert A. On Democracy. New Haven, CT: Yale University Press, 1998.
Darden, Keith, and Anna Grzymala-Busse. ‘The Great Divide: Literacy, Nationalism, and
the Communist Collapse’. World Politics 59, no. 1 (2006): 83–115.
Ebel, Robert, and Rajan Menon, eds. Energy and Conflict in Central Asia and the
Caucasus. Lanham, MD: Rowman & Littlefield Publishers, 2000.
Ekiert, Grzegorz, and Stephen E. Hanson, eds. Capitalism and Democracy in Central and
Eastern Europe. Assessing the Legacy of Communist Rule. Cambridge University
Press, 2003.
Elman, Colin. ‘Explanatory Typologies in Qualitative Studies of International Politics’.
International Organization 59, no. 2 (2005): 293–326.
Fish, M. Steven. Democracy Derailed in Russia: The Failure of Open Politics. Cambridge:
Cambridge University Press, 2005.
Fish, Steven M. ‘Democratization’s Requisites: The Postcommunist Experience’. PostSoviet
Affairs 14, no. 3 (1998): 212–47.
Fish, Steven. M. ‘The Determinants of Economic Reform in the Post-Communist World’.
East European Politics and Societies 12, no. 1 (1997): 31–78.
Fish, Steven M. ‘Stronger Legislatures, Stronger Democracies’. Journal of Democracy 17,
no. 1 (2006): 7–20.
Fish, Steven M., and Omar Choudhry. ‘Democratization and Economic Liberalization in
the Postcommunist World’. Comparative Political Studies 40, no. 3 (2007): 254–82.
Frye, Timothy. ‘A Politics of Institutional Choice: Post-Communist Presidencies’.
Comparative Political Studies 30, no. 5 (1997): 523–52.
George, Alexander, and Andrew Bennett. Case Study and Theory Development. MIT Press,
2005.
Gerring, John. Social Science Methodology: A Critical Framework. Cambridge: Cambridge
University Press, 2001.
Goertz, Gary. Social Science Concepts: A User’s Guide. Princeton, NJ: Princeton
University Press, 2006.
320 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
Hale, Henry E. ‘Regime Cycles: Democracy, Autocracy, and Revolution in Post-Soviet
Eurasia’. World Politics 58, no. 1 (2005): 133–65.
Ishiyama, John, and Matthew Velten. ‘Presidential Power and Democratic Development
in Post-Communist Politics’. Communist and Post-Communist Studies 31, no. 3
(1998): 217–33.
Janos, Andrew C. ‘Continuity and Change in Eastern-Europe: Strategies of Post-communist
Politics’. East European Politics and Societies 8, no. 1 (1993): 1–31.
Kim, Younkyoo. The Resource Curse in a Post-communist Regime: Russia in Comparative
Perspective. Aldershot: Ashgate, 2003.
Kitschelt, Herbert, Zdenka Mansfeldova, Radoslaw Markowski, and Gabor Toka.
Post-communist Party Systems. Competition, Representation, and Inter-Party
Cooperation. Cambridge: Cambridge University Press, 1999.
Kitschelt, Herbert. ‘Accounting for Outcomes of Post-Communist Regime Change: Causal
Depth or Shallowness in Rival Explanations?’ Paper presented at the annual meeting
of the American Political Science Association, Atlanta, 1999.
Kitschelt, Herbert. ‘Accounting for Postcommunist Regime Diversity. What Counts as a
Good Cause?’. In Capitalism and Democracy in Central and Eastern Europe.
Assessing the Legacy of Communist Rule, eds. Grzegorz Ekiert, Stephen E. Hanson,
49–88. Cambridge: Cambridge University Press, 2003.
Kopstein, Jeffrey S., and David A. Reilly. ‘Geographic Diffusion and the Transformation
of the Postcommunist World’. World Politics 53, no. 1 (2000): 1–37.
Kurtz, Marcus J., and Andrew Barnes. ‘The Political Foundations of Post-Communist
Regimes: Marketization, Agrarian Legacies, or International Influences’.
Comparative Political Studies 35, no. 5 (2002): 524–53.
Lazarsfeld, P.F., and A.H. Barton. ‘Qualitative Measurement in the Social Sciences:
Classification, Typologies, and Indices’. In The Policy Sciences, eds. D. Lerner and
H.D. Lasswell, 155–92. Stanford: Stanford University Press, 1951.
McFaul, Michael. ‘The Fourth Wave of Democracy and Dictatorship. Noncooperative
Transitions in the Postcommunist World’. World Politics 54, no. 2 (2002): 212–44.
De Melo, Martha, Cevdet Denizer, and Alan Gelb. ‘Patterns of Transition from Plan to
Market’. The World Bank Economic Review 10, no. 3 (1996): 397–424.
Merkel, Wolfgang. ‘Embedded and Defective Democracies’. Democratization 11, no. 5
(2004): 33–58.
IMF. Guide on Resource Revenue Transparency. http://www.imf.org/external/pp/
longres.aspx?id¼4176. 20 April 2007.
O’Donnell, Guillermo, and Philippe C. Schmitter. Tentative Conclusions about Uncertain
Democracies. Baltimore: MD The Johns Hopkins University Press, 1986.
Pop-Eleches, Grigore. ‘Historical Legacies and Post-Communist Regime Change’. Journal
of Politics 69, no. 4 (2007): 908–26.
Ragin, Charles C. The Comparative Method. Moving Beyond Qualitative and Quantitative
Strategies. Berkeley: University of California Press, 1987.
Ragin, Charles, and John Sonnett. Between Complexity and Parsimony: Limited Diversity,
Counterfactual Cases, and Comparative Analysis. COMPASSS working paper 2004-
23, http://www.compasss.org/files/wpfiles/RaginSonnett2004.PDF
Rihoux, Benoit, and Charles Ragin. Configurational Comparative Analysis. London: Sage,
2008.
Roeder, Philip G. ‘Varieties of Post-Soviet Authoritarian Regimes’. Post-Soviet Affairs 10,
no. 1 (1994): 61–101.
Sabonis-Helf, Theresa. ‘The Rise of the Post-Soviet Petro-states: Energy Exports and
Domestic Governance in Turkmenistan and Kazakhstan’. In In the Tracks of
Tamerlane: Central Asia’s Path to the 21st Century, eds. David Burghart and
Theresa Sabonis-Helf, 361–76. Washington, DC: National Defense University, 2004.
Democratization 321
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009
Sartori, Giovanni. The Theory of Democracy Revisited. Chatham, NJ: Chatham House
Publishers, 1987.
Schneider, Carsten Q., and Claudius Wagemann. ‘Reducing Complexity in Quantitative
Comparative Analysis: Remote and Proximate Factors and the Consolidation of
Democracy’. European Journal of Political Research 48, no. 5 (2006): 751–86.
Vachudova, Milada A. Europe Undivided: Democracy, Leverage, and Integration after
Communism. Oxford: Oxford University Press, 2005.
Wagemann, Claudius, and Carsten Q. Schneider. ‘Standards of Good Practice in Qualitative
Comparative Analysis (QCA) and Fuzzy-Sets’. COMPASS Working Paper 2007-51,
http://www.compasss.org/file/wpfiles/WagemannSchneider2007.pdf
322 J. Møller and S.-E. Skaaning
DownloadedBy:[SwetsContentDistribution]At:07:5824April2009