24
Stateness first? Jørgen Møller and Svend-Erik Skaaning Department of Political Science, Aarhus University, Aarhus, Denmark (Received 30 October 2009; final version received 5 May 2010) A number of scholars have recently claimed that ‘stateness’ is a prerequisite for democracy. However, a large-N empirical appraisal of this new research agenda is still pending. In this article, we demonstrate that stateness – conceptualized using the twin attributes of the monopoly on the use of force within a sovereign territory and a basic agreement about citizenship – is to a large extent a necessary condition for the four democratic attributes of electoral rights, political liberties, the rule of law, and social rights. However, the analyses also show that stateness is especially critical for the latter two attributes whereas the former two are sometimes encountered in its absence. These results are robust and they fit well into the dominant writings on democratization, which emphasize that in the present ‘liberal hegemony’ democratic elections are often grafted onto weak states – but that the rule of law and social rights are much more intimately wedded to structural constraints such as stateness. Keywords: stateness; necessity; electoral rights; political liberties; the rule of law; social rights; hierarchy Introduction ‘Democracy is a form of governance of a modern state. Thus, without a state, no modern democracy is possible’. With these words, Juan J. Linz and Alfred Stepan activated a prominent debate within the study of regime change: the debate about the relationship between what they termed ‘stateness’ and democ- racy. 1 Linz and Stepan emphasize two defining properties of stateness. First, that a sovereign state holding a monopoly on the use of force within the territory exists. Secondly, that the various groups in society have reached an agreement about who has the right to citizenship in this state. In this guise, stateness is expli- citly construed as a prerequisite for democracy by Linz and Stepan. However, modern democratic theory has – until recently – not devoted much attention to this issue. Such is the case because the preeminent democratic theorists of the twentieth century were preoccupied with properties pertaining to the political ISSN 1351-0347 print/ISSN 1743-890X online # 2011 Taylor & Francis DOI: 10.1080/13510347.2011.532607 http://www.informaworld.com Corresponding author. Email: [email protected] Democratization Vol. 18, No. 1, February 2011, 1–24

Stateness first?

Embed Size (px)

Citation preview

Page 1: Stateness first?

Stateness first?

Jørgen Møller∗ and Svend-Erik Skaaning

Department of Political Science, Aarhus University, Aarhus, Denmark

(Received 30 October 2009; final version received 5 May 2010)

A number of scholars have recently claimed that ‘stateness’ is a prerequisite fordemocracy. However, a large-N empirical appraisal of this new researchagenda is still pending. In this article, we demonstrate that stateness –conceptualized using the twin attributes of the monopoly on the use of forcewithin a sovereign territory and a basic agreement about citizenship – is toa large extent a necessary condition for the four democratic attributes ofelectoral rights, political liberties, the rule of law, and social rights.However, the analyses also show that stateness is especially critical for thelatter two attributes whereas the former two are sometimes encountered inits absence. These results are robust and they fit well into the dominantwritings on democratization, which emphasize that in the present ‘liberalhegemony’ democratic elections are often grafted onto weak states – butthat the rule of law and social rights are much more intimately wedded tostructural constraints such as stateness.

Keywords: stateness; necessity; electoral rights; political liberties; the rule oflaw; social rights; hierarchy

Introduction

‘Democracy is a form of governance of a modern state. Thus, without a state, nomodern democracy is possible’. With these words, Juan J. Linz and AlfredStepan activated a prominent debate within the study of regime change: thedebate about the relationship between what they termed ‘stateness’ and democ-racy.1 Linz and Stepan emphasize two defining properties of stateness. First, thata sovereign state holding a monopoly on the use of force within the territoryexists. Secondly, that the various groups in society have reached an agreementabout who has the right to citizenship in this state. In this guise, stateness is expli-citly construed as a prerequisite for democracy by Linz and Stepan.

However, modern democratic theory has – until recently – not devoted muchattention to this issue. Such is the case because the preeminent democratic theoristsof the twentieth century were preoccupied with properties pertaining to the political

ISSN 1351-0347 print/ISSN 1743-890X online# 2011 Taylor & Francis

DOI: 10.1080/13510347.2011.532607

http://www.informaworld.com

∗Corresponding author. Email: [email protected]

DemocratizationVol. 18, No. 1, February 2011, 1–24

Page 2: Stateness first?

regime. Schumpeter thus famously defined democracy as a Modus Procedendi ‘inwhich individuals acquire the power to decide by means of a competitive strugglefor people’s vote’.2 And in his influential definition of polyarchy, Dahl basicallyaugmented this definition by adding the political (and civil) liberties of freedomof association and expression to Schumpeter’s institutional arrangement.3

Arguably, until the onset of the so-called ‘third wave of democratization’, sucha parsimonious emphasis on the political regime was adequate to distinguishdemocracies from non-democracies. But as Guillermo O’Donnell has forcefullyargued that is no longer the case.4 More particularly, O’Donnell points out thatthe Schumpeterian and Dahlian focus on the political regime is based on certainimplicit premises anchored in the historical Western route to democracy. In gist,the constitutional infrastructure captured by the German phrase of the Rechtsstaatwas in place before the franchise was extended in the nineteenth and twentiethcentury.5 This was what T.H. Marshall observed with his famous tripartite divisionof citizenship rights into eighteenth-century civil rights, nineteenth-century politi-cal rights, and twentieth-century social rights.6

The Rechtsstaat was not the only stopover on the Western route to democracy,however. There was also a basic agreement about the borders of the state’s politicalcommunity – and this brings us back to the issue of ‘stateness’. As described byEugen Weber,7 Charles Tilly,8 and Martin van Creveld,9 there was absolutelynothing ‘democratic’ – or liberal – in the manner in which this challenge wastackled in Western Europe. Yet once solved, the basis for a well-functioningdemocracy was in place.

Not so in the developing world today. Under the current ‘liberal hegemony’,10

democratic elections are sometimes grafted onto states that are close to legalfictions and the rudimentary state apparatuses of which are thoroughly illiberal,illegitimate, and ineffective. According to Linz and Stepan, the path to democracyproper is strewn with obstacles in this situation, indeed possibly even shut. Overthe latest decade, this insight has gained a strong foothold in the study of regimechange.11 However, a systematic, large-N appraisal of the issue is still pending.In this article, we set out to deliver such an appraisal. To slightly retouchFukuyama,12 and with implicit reference to the attributes of democracy elucidatedbelow, our research question can be summed up thus: Stateness first?

Stateness, the Rechtsstaat, and democracy

One year after having described the concept of stateness with Stepan, Linz formu-lated the dictum ‘no state, no Rechtsstaat, no democracy’.13 The difference betweenthis formulation and the one quoted at the outset of this article goes to show thatsome preliminary conceptual spadework is pertinent. Generally speaking, the twoconcepts of stateness and the Rechtsstaat (or the rule of law) are often conflatedin the literature. Such is the case because much of the present literature basicallytreats stateness in terms of the closely related but broader concept of state capacity.We hold that this is problematical as state capacity also covers the ability to

2 J. Møller and S.-E. Skaaning

Page 3: Stateness first?

implement policies,14 something that obviously goes beyond the monopoly on theuse of force within a sovereign territory and the basic agreement about citizenshipemphasized by Linz and Stepan.15 Furthermore, and this brings us to the core con-ceptual problem, when construed as state capacity, the intension of the concept ofstateness is likely to overlap with that of the concept of the rule of law as thelatter – by definition – implies a certain level of administrative capacity.

Such an overlap is unfortunate because we risk conflating stateness with aspectsof liberal democracy, to some extent rendering the relationship between the two trueby definition. This point can be elucidated by revisiting O’Donnell. Based on theobservation about the historical Western route to democracy – reported in the intro-duction – he argues that democratic theory needs to go beyond the bare-bones focuson electoral rights and political liberties to embed an appreciation of the state as ‘alegal system that enacts and backs the universalistic and inclusive assignment ofthese rights and obligations’.16 This can be understood as a call for a ‘thicker’ defi-nition of democracy, including a conceptual slide from the regime to attributes ofthe state.17 Drawing on O’Donnell,18 we consider the rule of law to be defined inthe following way: the legal system upholds political and civil rights for the wholepopulation, and all public and private agents are subject to appropriate, legally estab-lished controls of the lawfulness of their acts, that is, no one is de legibus solutus.

Critically, this focus on the state’s legal system is not to be confused with ‘state-ness’ in Linz and Stepan’s sense; the intensions of the two concepts simply differ.Consequently, whereas the rule of law can be added as a defining attribute ofdemocracy by descending the ladder of abstraction, stateness is conceptuallydistinct from democracy; it is a prerequisite at most, not a defining attribute.

More generally, besides the core concept of stateness, we distinguish betweenfour attributes of democracy, viz. electoral rights, political liberties, the rule of law,and social rights. This provides a pivot for a conceptual distinction between‘thinner’ and ‘thicker’ concepts of democracy, as illustrated in Figure 1.

Electoral rights we derive from Schumpeter’s classic definition.19 Political lib-erties are anchored in Dahl’s elaboration (of polyarchy).20 The rule of law weembed with reference to O’Donnell’s call for augmenting the Dahlian definition.21

Finally, social rights are included with reference to Marshall’s seminal distinctionbetween different kinds of citizenship rights.22

In the empirical analyses, we test the relationship between stateness and thesefour attributes directly, rather than testing the relationship between stateness andthe four types of democracy. The reason for doing so is that this conceptual exercisein itself – due to the use of a ‘classical categorization’23 – implies the existence ofan empirical hierarchy, thus biasing the results in favour of our second hypothesisformulated below. Such is the case because the inverse relationship between theintension and the extension of a classical concept means that as we descend theladder of abstraction and add attributes, ever fewer empirical referents will be sub-sumed by a given concept.24 This is illustrated in Figure 1 above as the logic is thatthere will be more electoral democracies than polyarchies, more polyarchies thanliberal democracies and more liberal democracies than social democracies – for the

Democratization 3

Page 4: Stateness first?

simple reason that each additional category is created by augmenting the definition.Obviously, this makes for a hierarchical relationship between stateness and theever-thicker types of democracy as still fewer cases make it across the bar, irrespec-tive of their status on stateness. That, however, is not the case when employing thefour attributes on their own as neither subsumes any part of the others with regardto the intension of the concept.

One further point needs to be raised here. The inclusion of the attribute of socialrights might be said to transgress the limits imposed by the Schumpeterian ‘back-ground concept’ of democracy as a political method. The first thing we would liketo note here is that, as explained in the preceding paragraph, we test the relationshipbetween stateness and each of the four attributes in turn. Consequently, we do notemploy ‘social democracy’ – only social rights – as explanandum in the empiricalanalysis. Social rights can at the very least be understood as an accompanying attri-bute of democracy,25 and many scholars have gone further and virtually included itas a defining attribute.26 As we wish to carry out an adequate appraisal of theliterature, we include it too.

Figure 1. Descending the ladder of abstraction to construct types of democracy.

4 J. Møller and S.-E. Skaaning

Page 5: Stateness first?

The research question can now be spelled out more clearly. As implied by thewording ‘stateness first’, it can be understood in terms of necessary causation.Relating this to the four attributes of democracy, we may reformulate the questionthus: To what extent is stateness a necessary condition for electoral rights, politicalliberties, the rule of law, and social rights, respectively?

True by definition?

But is stateness – in the guise of a monopoly on the use of force within the territoryand a basic agreement about the demos – a logical or an empirical prerequisite forthe presence of the democratic attributes listed above? That is, is the relationshipnecessary by definition or because it reflects a causal relationship? And, insofaras it is only an empirical precondition, why might it function as such to a differentextent with regard to the four democratic attributes?

To answer these questions, it is helpful to revisit Robert A. Dahl’s famousobservation that ‘the criteria of the democratic process presuppose the rightfulnessof the unit itself’.27 Echoing – and referring to – this observation, Linz and Stepanunderline that stateness is ‘logically prior to the creation of democratic insti-tutions’.28 Elkins and Sides explain this as follows:

The issue is one of consent. While democracy requires that citizens accept the legiti-macy of the elected leaders and rules that put them there, it also requires, morefundamentally, that citizens respect the prerogatives and boundaries of the statethat these leaders govern.29

As mentioned in the introduction, we take our cue from Linz and Stepan. But wediffer from them on this particular point. We argue that stateness cannot simply beconstructed as a logical fact prior to the democratic process. Recall in this connectionthat the definitions of the attributes spelled out above deliberately ensure that theintension of stateness does not overlap with that of the four democratic attributes.

Notwithstanding this crisp conceptual distinction, Linz and Stepan’s formu-lation might be said to make sense if two conditions were satisfied. First, if state-ness is operationalized in dichotomous terms (‘either-or’)30 and, secondly, if eachcountry operates in a closed system with no external input. Neither is the case here,however.

The first of these conditions does not obtain as we test the necessity of statenessfor the four democratic attributes at different levels of infringement. While it mightbe valid to assert that a particular threshold of stateness must be crossed to bringabout electoral rights, it would be a very different matter to argue that electoralrights can – logically – never exhibit higher levels of defects than stateness.

The second of these conditions is arguably not met empirically. We proposethat in today’s world, external pressure has an independent effect on the statusof some of the democratic attributes, even when stateness is accounted for. Thispoint is elaborated below.

Democratization 5

Page 6: Stateness first?

To sum up: The extent to which stateness, measured on different levels ofinfringement, obtains as a necessary condition for electoral rights, political liber-ties, the rule of law, and social rights is not a matter that can be solved logically.Rather, it is an empirical problem.

Stateness and the attributes of democracy: two propositions

Is there a systematic difference in the extent to which we should expect stateness towork as a necessary condition for each of these four attributes? Based on recentwritings on regime change, our answer to this question is in the affirmative. Anumber of scholars have claimed that the present ‘liberal hegemony’ favours thespread of democratic elections and political liberties in developing countries.31

However, they are normally keen to add that the development of other aspectsof modern democracy – the rule of law and social rights – is much more intimatelywedded to structural constraints.32

Stateness is of course an ultimate example of such structural constraints. Ittherefore makes sense to probe whether stateness bears upon the two-fold hierar-chy – derived from the literature – between electoral rights and political libertieson the one hand and the rule of law and social rights on the other hand.33 Based onLinz and Stepan’s arguments about the critical importance of stateness for democ-racy and the writings emphasizing the consequences of the present liberal hege-mony we suggest two propositions:

H1: In developing and transformation countries of the contemporary era, stateness isto a large extent a necessary condition for obtaining each of the four attributes ofdemocracy, that is, electoral rights, political liberties, the rule of law, and social rights.

H2: In developing and transformation countries of the contemporary era, the necess-ary status of stateness for the attributes of democracy takes the form of a particularhierarchical pattern: it is most critical for social rights and the rule of law and leastcritical for political liberties and electoral rights.

The fact that the problem has been formulated in terms of necessary causationmight lead some to dispute the H1-qualifier ‘to a large extent’ and, ipso facto, thenotion of a hierarchical pattern stated in H2. Both of these patterns are of course ren-dered meaningless if we consider necessity in rigid deterministic terms, that is,when not allowing for any exceptions. But perfectly consistent relationships areextremely rare in the social sciences. As Ragin points out this is so because thereis ‘a great deal of randomness and chance in human affairs, and many flaws anderrors in the efforts of social scientists to record and decipher social life’.34

Furthermore, and as already hinted, strong theoretical reasons can be offered toexplain why stateness in the present era of ‘liberal hegemony’ is more critical forthe rule of law and social rights than for electoral rights and political liberties. Withdue reference to Ragin35 and Goertz and Starr,36 we argue that the assessment ofstateness as a necessary condition makes sense in probabilistic terms. Also, and

6 J. Møller and S.-E. Skaaning

Page 7: Stateness first?

as already mentioned, the assessment of necessity need not be restricted to a dichot-omous (presence/absence) perspective. As a necessary relationship is basically aset theoretical subset relationship, variations by level are allowed for. Therefore,the degree of necessity can be assessed by examining the extent to which member-ship in the outcome is consistently less than or equal to membership in the cause.This is exactly what we do in our empirical analyses.

The scope of the inquiry

In the analyses below we take stock of ‘transformation and developing countries’in the year 2007. This is the scope of the dataset which we use to test the twohypotheses, viz., the Bertelsmann Transformation Index 2008 (BTI). BTI is anambitious attempt to create an index of politico-economic transformation in non-OECD countries. As it is also a relatively new dataset – first produced in 2005– a description of the data-generation process is warranted. Each of the 125countries included in the 2008 version (covering 2007) are initially treated in acountry report, which analyses the country’s performance on a total of 17 criteria(broken down into 49 questions). The actual index is based on a questionnaire onthese 49 questions, each of which are scored on a scale ranging from 10 (best) to 1(worst) by a country expert. The questions are grouped into various clusters, suchas Stateness, Political Participation, and the Rule of Law.

The numerical ratings of the country expert are reviewed three times. First, anadditional country expert scores the country independently on the 49 questions.Two regional experts then discuss these ratings and agree on ratings that reflectthe differences among countries of the same region. The regional coordinatorsand the BTI team subsequently convene and review ratings across regions to cali-brate the scores. Insofar as a calibrated score differs significantly from thatsuggested by the first expert, he/she is consulted. Final rating decisions,however, are made by the BTI board.

As laid out in the subsequent discussion, the BTI is the only dataset well-suitedfor achieving agreement between the definition and the measurement of both state-ness and the four attributes of democracy. However, based on the standards pre-sented by Munck and Verkuilen,37 we also argue that the measures inthemselves have a competitive edge over those presented by available datasets,such as the Worldwide Governance Indicators38 and the Freedom in the WorldSurvey.39 The BTI thus observes a high standard with respect to its attention todetail (many components, a fine-grained scale, and a relatively precise coding-manual) and the thoroughness with which these are treated. What is more, itscoherent framework and reliance on experts makes for measurement reliabilityand validity. That said, the BTI is not without problems as, inter alia, no formalinter-coder reliability test is carried out, some of the indicators, such as the onenamed civil rights, collapses a number of distinct aspects, the cross-temporal cov-erage is limited, and the same goes for the country-coverage (as mentioned, the BTI2008 only includes 125 non-OECD countries).40

Democratization 7

Page 8: Stateness first?

The limited country coverage does, however, not pose a problem for our analy-sis as our hypotheses are not formulated to cover the developed countries. Thiscross-spatial scope condition can be justified both theoretically and pragmatically.In gist, the interesting variation on both the explanandum and explanans is to befound in the developing world. Such is the case because virtually all the developed(OECD) countries score high on stateness and exhibit the bounded whole of elec-toral rights, political liberties, the rule of law, and social rights. This in no way goesagainst the findings of this paper. But we need variation on stateness (and the fourattributes of democracy) to carry out an empirical examination of our propositions– and this variation is to be found outside the developed Western world. Nonethe-less, we have attempted to go beyond the scope condition by including all countriesin one of the robustness tests at the end of the paper.

The temporal scope condition of liberal hegemony is also warranted both theor-etically and pragmatically. Only after the end of the Cold War has the externalpressure to democratize become so salient that electoral rights and the associatedpolitical liberties are likely to be encountered in even some countries that do nothave the attribute of stateness. But, critically, the rule of law and social rightsare still extremely difficult to achieve in this situation. These are of course thetwin dynamics theoretically making for the predicted hierarchy (H2).

Still, why not cover a longer time span in the analysis to test the merit of thistemporal scope condition? Obviously, it would be interesting to trace the patternback in time to the onset of liberal hegemony – and to the antecedent period forthat matter. Yet this is simply not possible in the analysis proper. Though cross-temporal measurements of electoral rights and political liberties exist, adequatemeasures of stateness, social rights, and, albeit to a lesser extent, the rule of laware not readily available. Only with the advent of the BTI has it become possibleto measure the levels of infringements on all five attributes for a large number ofcountries. For this reason, up until the robustness tests the only other year weanalyse is 2005, which is also included in the BTI.

Recalling the conceptual definitions of this paper, the BTI contains twelvesub-components that are relevant for our analyses. We have distributed themunder the five respective headings as illustrated in Table 1 (see Appendix 1,Table A1 for the detailed description of the indicators). To the extent thatscores from more than one subcategory are used to capture an attribute, we con-sider them to be mutually constitutive. Consequently, rather than using theaverage, we employ a minimum score procedure to aggregate them, as rec-ommended by Gary Goertz.41

On methodology

Harking back to the research question and the two hypotheses, the empirical analy-sis presupposes methods capable of handling causal conditions in terms of neces-sity. This might seem an obvious point but it is one that is often ignored in actualempirical analyses. In fact, it is not uncommon to encounter studies that formulate

8 J. Møller and S.-E. Skaaning

Page 9: Stateness first?

hypotheses in terms of necessity and sufficiency only to test these using standardcorrelational methods.42 This procedure is inappropriate for a very simple reason:to suggest that stateness is necessary for democracy is different from suggestingthat the higher the level of stateness, the higher the level of democracy.

More generally, researchers should be conscious about the logical forms of therelationships they are interested in and use the appropriate methods rather than the‘industry standard’.43 In what follows, we use two methods that are both capable oftesting relationships stated in terms of necessity.

We first create a simple conceptual typology that makes for distinguishing sys-tematically between stateness on the one hand and the respective attributes of elec-toral rights, political liberties, the rule of law, and social rights on the other hand.Such typologies present a device for assessing explanatory variables in terms ofnecessary and sufficient causation.44 Ordering the transformation and developingcountries in this typology therefore allows us to carry out a first test of whether theempirical distributions support the hypotheses. To evaluate the strength of theempirical fit of the hypotheses, we use the proportion of referents classified inaccordance with the theoretical expectations.

Subsequently, we use fuzzy set Qualitative Comparative Analysis (fsQCA) toexamine whether the typological findings are robust when employing a morerefined measure of consistency that is capable of appreciating the differencebetween large and small inconsistencies and thus the differences between thefine-grained scores on the attributes. We here adopt Ragin’s yardstick sayingthat a strong hierarchical (subset) relationship should show consistency levelsabout 0.85 or more.45 As the analytical logic of the two applied methods is basi-cally of the same ilk, this criterion also applies to the typological results.

A typological analysis

The typological analysis simply relates stateness to each of the four attributes ofdemocracy. The detailed BTI scores for each component range from 1 to 10,with 10 indicating the highest level of accomplishment. We use the four-fold

Table 1. Linking the BTI sub-components to the five attributes.

Stateness Electoral rights Political liberties Rule of law Social rights

(i) Monopoly onthe use offorce

(i) Free elections (i) Association/assemblyrights

(i) Separation ofpowers

(i) Socialsafety nets

(ii) Citizenshipagreement

(ii) Democratsrule

(ii) Freedom ofexpression

(ii) Independentjudiciary

(ii) Equal

opportunity(iii) Abuse of

officeprosecuted

(iv) Civil rights

Democratization 9

Page 10: Stateness first?

hierarchy of linguistic qualifiers in the codebook that guide the expert assessmentsto distinguish between four classes on each attribute in the following way: ‘nodefect’ (9–10), ‘moderate defect’ (6–8), ‘severe defect’ (3–5), and ‘extremedefect’ (1–2).

Consequently, each test of the relationship between stateness and the fourattributes can be construed as a property space consisting of 16 (42) types. Howdo four such simple property spaces make for a test of the hypotheses? H1 requiresthat the 125 countries included in the BTI numbers for 2007 to a large extent clusterin the 10 shaded types situated along or below the top-right to bottom-left diagonal(the antidiagonal) of each of the four property spaces. If this is the case, then itequals saying that stateness – on each of the four levels of infringement – isbasically a necessary condition for the attributes of democracy.

Yet there should be more to it than that. The hierarchical logic of H2 impliesthat though absolutely high across the board, the relative differences should findexpression in consistency levels that are highest for social rights and the rule oflaw and lowest for political liberties and electoral rights. How do the empiricalfacts fit with these expectations?

As illustrated in Table 2, the empirical ordering provides quite strong generalsupport for the hypotheses. Recall that only cases situated outside the shaded cellsdisconfirm the claim of a necessary relationship. A random distribution would haveproduced a consistency of 0.63 (10/16) (and no less than 46 empirical misfits outof the 125 countries). Bearing this in mind, the goodness of fit of H1 is fairly high inthe twin cases of electoral rights, viz. 0.84 (105/125), and political liberties, viz.

Table 2. Typological ordering of countries, 2007.

Stateness

Extremedefects

Severedefects

Moderatedefects

Nodefects

Electoral rights No defects 0 0 12 17Moderate defects 0 5 26 7

Consistency ¼ 0.84 Severe defects 3 4 4 1Extreme defects 2 8 28 8

Political liberties No defects 0 1 6 12Moderate defects 1 7 32 12

Consistency ¼ 0.86 Severe defects 3 7 21 6Extreme defects 1 2 11 3

Rule of law No defects 0 0 1 5Moderate defects 0 1 12 12

Consistency ¼ 0.98 Severe defects 1 11 43 11Extreme defects 4 5 14 5

Social rights No defects 0 0 0 5Moderate defects 0 1 14 18

Consistency ¼ 0.98 Severe defects 1 11 48 9Extreme defects 4 5 8 2

10 J. Møller and S.-E. Skaaning

Page 11: Stateness first?

0.86 (107/125). And it is near to perfect in the twin cases of the rule of law, viz.0.98 (122/125), and social rights, viz. 0.98 (123/125).

Using the somewhat arbitrary benchmark of 0.85, it can of course be debatedwhether stateness is indeed a necessary condition for electoral rights and politicalliberties. This obviously depends on the threshold for such necessity and, in allevents, even electoral rights only misses the 0.85 threshold by a hair’s breadth.Furthermore, no such doubts can be entertained with regards to the rule of lawand social rights. Out of 125 countries, only three and two, respectively, showhigher scores on these attributes than on stateness. The difference between thetwo groups of attributes is thus both precipitous and in line with the pattern postu-lated in H2.

Replacing the numbers with proper names, the five said aberrations are Afgha-nistan, Bosnia, and Estonia (on the rule of law) and Iraq and Sri Lanka (regardingsocial rights). All of these exceptions are minor in that they are situated in a type(cell) adjacent to those making for necessity. However, they are also very muchunanticipated because they break with the manifest pattern regarding the twomore demanding attributes of the rule of law and social rights and, ipso facto,with H2. One obvious analytical observation suggests itself here. All thesecountries have been embroiled in dispute – in four of the cases taking the formof open conflict – on ethnic basis and three are currently under international super-vision (Afghanistan, Bosnia, and Iraq). This is likely to make for idiosyncraticassociations between stateness and the four attributes.

With regard to electoral rights and political liberties, 20 and 18 aberrations canbe identified. Once again, most of these exceptions are situated in cells adjacent tothose making for necessity. Scanning the identity of these misfits – which aremisfits to H1 only – no obvious or interesting patterns can be identified. Thecases constitute a much more heterogeneous bunch than those breaking with H2,although the five countries just mentioned are also included here. This finding initself goes someway to support H2 as it shows that idiosyncratic problems ofstateness are not a precondition for breaking the waves with respect to the twoless-demanding democratic attributes.

Consequently, the empirical data sustain H2 both quantitatively (regarding thegeneral fit) and qualitatively (regarding the face validity of the exceptions). Notice,finally, that though largely necessary for all the democratic attributes, stateness isby no means a sufficient condition for either of these. In fact, many countries fall inthe shaded cells below the antidiagonal – for example, exhibiting no defects onstateness but moderate, severe or extreme defects on the four attributes – thusdisconfirming any claim of a sufficiency relationship.

A fuzzy-set analysis

To investigate whether a similar hierarchical pattern exists when using the detailedBTI scores ranging from 1 to 10, we employ fsQCA. This means that – vis-a-visthe typological analysis using the simple property spaces – information is not lost

Democratization 11

Page 12: Stateness first?

through collapsing values (for example, 6, 7, and 8 into the status of moderatedefect). By applying the formula provided by Ragin to calculate theconsistency of necessary conditions it is thus possible to carry out what maybest be understood as a refined robustness test of the typological findings.46

The results – illustrated in Table 3 – show that the prior findings are not over-turned. Indeed, they are very much supported, as the levels of consistency clearlypass the benchmark criterion of 0.85 this time around. In fact, in all instances theyare higher than the equivalent consistency levels that emerged in the typologicalanalysis. H1 is thus clearly corroborated.

To illustrate the nature of this systematic pattern – and the logic of the analysis– visually, we have included a plot of the stateness scores against the rule of lawscores in Figure 2. This plot can be construed as an attempt to assess Linz’dictum quoted above: ‘no state, no Rechtsstaat, no democracy’. Notice that vir-tually all cases fall below the antidiagonal and that they are distributed throughoutthis area. This indicates, first, that stateness is a necessary but not a sufficient

Table 3. Assessing the degree of necessity by means of fsQCA, 2007.

Electoral rights Political liberties Rule of law Social rights

Level of consistency 0.92 0.90 0.99 0.99

Figure 2. Scatter (XY) plot of stateness scores and the rule of law scores, 2007.

12 J. Møller and S.-E. Skaaning

Page 13: Stateness first?

condition for the rule of law, secondly, that the relationship cannot be modelledusing standard statistical techniques, and, thirdly, that stateness is not a ‘trivial’ pre-requisite. The latter would be the case if all cases were placed at the right hand sideof the figure, meaning that they all scored high on stateness.

The notion of an empirical hierarchy, as implied by H2, is also sustained by thefsQCA. Such is the case because the hierarchical pattern between stateness on theone hand and the four attributes on the other hand is characterized by differentconsistency levels. Once again, the consistency levels associated with the twofirst attributes, that is, electoral rights (0.92) and political liberties (0.90), arebarely distinguishable. And once again, this is even more so for the two otherattributes (the rule of law and social rights); the strength of the necessary relation-ship is almost perfect (0.99) in both cases.

Yet between these two groups the difference is precipitous. All of this goes toshow, first, that stateness is very close to being a necessary condition for all fourattributes but, secondly, that it is more of a necessity for social rights and therule of law than for political liberties and electoral rights.

Some additional robustness tests

To further underpin these conclusions, we have carried out two rounds of robust-ness tests. The first round is based on the BTI only, the second one uses alternativedatasets.

The BTI robust test can be summed up in four steps. First, we have rerun theanalysis applying the standard – but in our view flawed – aggregation procedure,which consists in using the simple mean of the indicators for both stateness and thefour attributes. Secondly, we have augmented the definition and operationalizationof stateness with an additional attribute, derived from the BTI, which measures towhat extent basic administrative structures exist. This can be understood as anattempt to make the concept and measurement more tangential to the concept ofstate capacity. Thirdly, we have relocated the sub-component of civil rights fromthe third attribute (the rule of law) to the second attribute (political liberties),which makes for construing the former as horizontal accountability and the latteras civil liberties.

As illustrated in Table 4, these modifications further substantiate the generalpicture conjured up by this study, viz. that modern democracy to a large extent pre-supposes stateness, yet that this prerequisite is more critical for the rule of law andsocial rights than for political liberties and electoral rights.

Finally, we have carried out a robustness-test of the original analysis with theonly other year for which the adequate data exist. This is 2005, in which the BTIincludes data for 119 developing and transformation countries. The consistency for2005 is reported in the last row of Table 4. Once again, a similar pattern appears.

More interesting is the additional set of robustness tests which are a product ofembracing alternative datasets. As already argued, no additional dataset is satisfac-tory when it comes to measuring either stateness or the four attributes of

Democratization 13

Page 14: Stateness first?

democracy. However, the sub-component scores of the Freedom House’s so-calledFreedom in the World Survey that were first made public in 2006 (covering 2005)make it possible to derive relatively valid measures of three of the four attributes ofdemocracy, viz. electoral rights, political liberties, and the rule of law.47 As thedeveloped countries are also included in the Freedom House’s Freedom in theWorld Survey (FH) this means that we can go beyond the cross-spatial scopecondition insofar as a more encompassing measure of stateness can be found.

Considering our definition of stateness, no such alternative dataset currentlyexists. As a lesser evil, we have employed the so-called Failed States Index(FSI), provided by The Fund for Peace, which is sometimes referred to as aproxy for stateness in empirical analyses. We would like to emphasize that wedo not find this a valid representation of our concept since the conceptualizationis not in agreement with our definition and as the methodology behind the FSI isflawed.48 These validity problems mean that the robustness tests that incorporatethe FSI should not be referred in their own right. That said, for the purposes of atentative check it might still have some value.49

On this basis, four additional robustness tests are carried out, as illustrated inTable 5. In the first test, the FH measure of the democratic attributes has been substi-tuted for the BTI measure. As BTI is still employed to measure stateness, only thedeveloping countries are included here. In the second test, the FSI measure of state-ness is combined with the BTI measure of the four democratic attributes, again cover-ing the developing world only. In the third test, the FSI measure of stateness iscombined with the FH measure of the democratic attributes, but the analysis isdeliberately restricted to the same scope as the BTI. Finally, in the fourth analysis,the same two alternative measures are used to incorporate the developed countriesas well.

The results practically interpret themselves. To the extent that the tests arevalid, the findings of the analysis proper emerge as robust. The only minorhiccup is that the consistency of stateness as a necessary condition is marginallyhigher for electoral right than for political liberties – which may be said to goagainst the most pure version of the expected hierarchy. However, the important

Table 4. Assessing the degree of necessity by means of fsQCA, using the modifiedmeasurements.

Electoralrights

Politicalliberties

Rule oflaw

Socialrights

Level of consistency (alteration:aggregation procedure)

0.96 0.95 0.99 1.00

Level of consistency (alteration:measurement of stateness)

0.85 0.84 0.98 0.99

Level of consistency (alteration:relocation of civil rights)

0.92 0.95 0.99 0.99

Level of consistency (alteration:numbers for 2005)

0.91 0.89 0.94 0.98

14 J. Møller and S.-E. Skaaning

Page 15: Stateness first?

part of the hierarchy, that the rule of law and social rights are in a league of theirown with respect to the consistency, comes out strongly.

Is the relationship spurious?

At first sight, the findings of this paper are based on a simple bivariate analysis.Some readers may find this problematical as any empirical establishment of caus-ality rests on the ceteris paribus clause, meaning that relevant third variables mustbe controlled for. In gist, it is necessary to examine whether including other expla-natory variables give us reasons to think that the relationship is spurious.

To confront this challenge, we first have to distinguish between antecedent andintervening factors. To reveal whether a relationship is spurious, only an examin-ation of potential antecedent causes is warranted. As Mahoney has emphasized inhis discussion of set theoretical relationships, the effect of X (as a necessary con-dition for Y) – in our case stateness as a necessary condition for the four demo-cratic attributes – is to be considered spurious if an antecedent condition, Z, isat the same time sufficient for X and necessary for Y.50

Herbert Kitschelt’s influential distinction between ‘deep’ and ‘proximate’ factorsalso bear upon the issue of spuriousness.51 To be an antecedent cause of stateness, afactor must be ‘deeper’ in Kitschelt’s sense, that is, situated further from the explanan-dum temporally. More proximate causes can be intervening at most. Importantly,stateness – though measured in the 2000s – is a very deep factor. It is stronglypath-dependent and therefore almost impervious to change in the short run. Indeed,it is deeper than most other factors which can be identified within the democratizationliterature, thereby limiting the number of relevant third variables considerably.

Only one important explanatory variable from the literature goes some way tofulfilling the criterion that we have formulated. That is the level of modernizationthat has often been hailed as a prerequisite of democracy and that has clearly inter-acted with, and possibly shaped, state- and nation-building over the latest centuries.

Table 5. Assessing the degree of necessity by means of fsQCA, using alternative measures,2007.

Electoralrights

Politicalliberties

Ruleof law

Socialrights

Level of consistency (BTI measure of stateness,FH measure of the democratic attributes)

0.94 0.94 0.99 –

Level of consistency (FSI measure of stateness,BTI measure of democratic attributes)

0.85 0.84 0.98 0.98

Level of consistency (FSI measure of stateness, FHmeasure of the democratic attributes,developing countries)

0.91 0.90 0.99 –

Level of consistency (FSI measure of stateness, FHmeasure of the democratic attributes, allcountries)

0.92 0.91 0.98 –

Democratization 15

Page 16: Stateness first?

To test this, we use the calibrated GNI/capita-figures included in the BTI to oper-ationalize the level of modernization.52

The results show that modernization easily passes the consistency criterion forbeing sufficient for stateness, with a level of 0.95. However, for two reasons thisfinding does not undermine the main conclusion. First, modernization is neithernecessary for electoral rights (0.55), political liberties (0.54), the rule of law(0.74), or – though closer to being so – social rights (0.81). Secondly, the directionof the causal relationship with stateness is not cut in stone. One could more convin-cingly argue that stateness is a necessary condition for modernization53 – or, alter-natively, that the two conditions are mutually reinforcing. Thus, the relationshipbetween stateness and the attributes of democracy does not appear spurious.

Conclusions

We set out to deliver an appraisal of the current debate – sparked by Linz andStepan over a decade ago – about the relationship between stateness and democ-racy. The conceptual premise of this empirical exercise was a simple one: that it isvital not to conflate stateness with attributes of democracy such as the rule of law.There is an intrinsic danger in doing so because stateness is often simply construedas state capacity in the literature. Zealously sticking with Linz and Stepan, we haveargued that stateness should instead be understood in terms of the existence of amonopoly of violence in a sovereign territory and an agreement about citizenshipwithin this polity. In this guise, stateness is not only distinct from democracy; it alsoworks as a ‘deep’ constraint on democratization.

This conceptual clearing operation provided the stepping stone for testing theextent to which stateness is a necessary condition for different attributes of democ-racy. Analysing contemporary developing and transformation countries in 2007(using the BTI 2008), we encountered some manifest empirical regularities.First, stateness is to a large extent a requisite for all of the four attributes of electoralrights, political liberties, the rule of law, and social rights. Secondly, the analyseslaid bare an empirical hierarchy in that stateness is most critical for social rightsand the rule of law – indicated by almost perfect necessary relationships – andleast critical for political liberties and electoral rights.

More generally, the empirical results supported our hypotheses, which werededuced from the dominant current writings on democratization and whichembedded the theoretical logic of stateness described by Linz and Stepan. Thelevel of stateness is simply imprinted in the body politic of contemporary develop-ing and transformation countries. This critical stopover on the Western route todemocracy is – it seems – equally critical in developing countries of the contem-porary era. For long this has been ignored because the problem of stateness hadbeen solved in the developed countries before the widening of the franchise. Butif we wish to expound the travails of democracy under the auspices of the thirdwave, then an understanding of this issue is essential.

That does not imply that stateness is – to reuse Linz and Stepan’s formulation– ‘logically prior’ to the democratic attributes. Our empirical analysis clearly

16 J. Møller and S.-E. Skaaning

Page 17: Stateness first?

showed that exceptions exist in the real world, at least when the relationship isinvestigated at different levels of infringement. More precisely, whereas only afew countries break with the logic of necessity regarding the attributes of therule of law and social rights quite some countries do so as regards electoralrights and political liberties.

Neither should this be taken to mean that stateness in itself makes for democ-racy. What our study has shown is basically that stateness is to a large extent anecessary condition for democracy to gain (and retain) foothold and for democracyto become ‘thicker’. It is, however, not a sufficient condition. On the contrary, ouranalyses shows that a sizeable number of countries do better on stateness than on thevarious attributes of democracy – it is only the opposite combinations that are rare.

What does this imply with respect to the current dynamics of democratization?The factors that influence the type or level of democracy are obviously many. Never-theless, the findings of this article strongly suggest that many countries are unable todevelop democracy because of severe shortcomings on the attribute of stateness. Tospell it out, the lack of a monopoly on the use of force in a sovereign territory and thelack of consensus about who holds legitimate citizenship within this unit rendersdemocratization virtually impossible. To the basic question, ‘Stateness First?’, ourgeneral answer is therefore ‘yes’ – but with a twist. It is more of a prerequisite forthe rule of law and social rights than for electoral rights and political liberties. A for-tiori, in the contemporary era of ‘liberal hegemony’ problems of stateness are morerelevant for the thicker types of democracy (liberal democracy and social democ-racy) than for the thinner types of democracy (electoral democracy and polyarchy).

Different ways of handling the problem of stateness have been offered in theliterature,54 including provisional shared sovereignty,55 international recognitionof de facto sovereign units, and reconstruction boundaries when needed.56 Eventhough we have not directly addressed this research agenda, our results certainlysupport its relevance.57

AcknowledgementsFor their comments and suggestions, we are indebted to Charles Ragin, Frank Fukuyama,Wolfgang Muno and, for Democratization, two anonymous reviewers. The usual disclaimerapplies.

Notes1. Linz and Stepan, Problems of Democratic Transition and Consolidation, 17.2. Schumpeter, Capitalism, Socialism and Democracy, 269.3. Dahl, Polyarchy.4. O’Donnell, Dissonances. See also Kraxberger, ‘Failed States’, 1056–7.5. Cf. Rose and Shin, ‘Democratization Backwards’.6. Marshall, Citizenship and Social Class.7. Weber, Peasants into Frenchmen.8. Tilly, Coercion, Capital and European States.9. Van Creveld, The Rise and Decline of the State.

10. Levitsky and Way, ‘The Rise of Competitive Authoritarianism’, 61.

Democratization 17

Page 18: Stateness first?

11. For example, Fukuyama, ‘The Imperative of State-Building’; Fukuyama, ‘“Stateness”First’; Fukuyama, ‘Liberalism versus State-Building’; Tilly, Democracy; Kraxberger,‘Failed States’; Bratton and Chang, ‘State Building and Democratization’; Rose andShin, ‘Democratization Backwards’.

12. Fukuyama, ‘“Stateness” First’.13. Linz, ‘Democracy Today’, 118.14. See for example Fukuyama, ‘“Stateness” First’; Evans, ‘The Eclipse of the State?’.15. See also Back and Hadenius, ‘Democracy and State Capacity’, 3.16. O’Donnell, ‘Democracy, Law, and Comparative Politics’, 18.17. Collier and Levitsky, ‘Democracy with Adjectives’, 445–6.18. O’Donnell, ‘Polyarchies and the (Un)rule of Law’; O’Donnell, ‘Human Development,

Human Rights, and Democracy’.19. Schumpeter, Capitalism, Socialism and Democracy. Our definition of the ‘electoral’

attribute is, however, somewhat more demanding than that of Schumpeter as we donot allow for the exclusion of significant parts of the population or illegitimate vetopowers such as the military. On this issue, see Møller and Skaaning, ‘Beyond theRadial Delusion’.

20. Dahl, Polyarchy.21. O’Donnell, ‘Democracy, Law, and Comparative Politics’.22. Marshall, Citizenship and Social Class.23. Cf. Collier and Mahon, ‘Conceptual “Stretching” Revisited’.24. Sartori, ‘Concept Misformation’.25. Cf. Sartori, ‘Guidelines for Concept Analysis’.26. For example, Dahl, Dilemmas of Pluralist Democracy; Dahl, A Preface to Economic

Democracy; Beetham, Democracy and Human Rights.27. Dahl, Democracy and Its Critics, 207.28. Linz and Stepan, Problems of Democratic Transition and Consolidation, 26.29. Elkins and Sides, Seeking Stateness, 1.30. Even if stateness was in fact construed in terms of one particular threshold, an empiri-

cal appraisal would still be preferable to a solely logical discussion.31. Nodia, ‘How Different Are Postcommunist Transitions?’, 22; Hadenius, Democracy’s

Victory and Crisis, 3; Burnell and Calvert, ‘The Resilience of Democracy’, 1; Levitskyand Way, ‘The Rise of Competitive Authoritarianism’.

32. See Grzymala-Busse and Luong, ‘Reconceptualizing the Post-Communist State’, 536;O’Donnell, ‘Democracy, Law, and Comparative Politics’; Sørensen, Democracy andDemocratization; Diamond, Developing Democracy. Furthermore, in a number ofrecent papers (Møller and Skaaning, ‘Beyond the Radial Delusion’; Møller and Skaan-ing, ‘Marshall Revisited’), we have demonstrated that in developing and transform-ation countries just such an empirical hierarchy among the highlighted attributes ofdemocracy currently exists. More particularly, the data show that electoral rights areat least as effective as (and often more effective than) political liberties which are atleast as effective as (and often more effective than) the rule of law which is at leastat least as effective as (and often more effective than) social rights.

33. A particular challenge must be addressed here. Bratton and Chang (‘State Buildingand Democratization’, 1061) argue ‘that there exists a self-reinforcing cycle inwhich state building and regime consolidation feed each other’. This implies that state-ness is not only a condition of democratization but also a product thereof. However,insofar as the present ‘liberal hegemony’ on the international scene has an asymme-trical impact on the four attributes, in particular by privileging electoral rights andcivil liberties over the rule of law and social rights, this would make for a hierarchicalordering even if such endogeneity is present to some degree.

34. Ragin, ‘Fuzzy-Set Analysis of Necessary Conditions’, 180.

18 J. Møller and S.-E. Skaaning

Page 19: Stateness first?

35. Ibid.; Ragin, ‘Set Relations in Social Research’.36. Goertz and Starr, ‘Introduction’.37. Munck and Verkuilen, ‘Conceptualizing and Measuring Democracy’.38. Kaufman, Kray, and Mastruzzi, Governance Matters VIII.39. Freedom House, Freedom in the World.40. Skaaning, ‘Measuring Civil Liberty’; Skaaning, ‘Measuring the Rule of Law’.41. Goertz, Social Science Concepts; see also Bowman, Lehoucq, and Mahoney, ‘Measur-

ing Political Democracy’. Notice that the civil rights component covers the issues ofequality before the law, due process, and personal integrity rights which is why wehave situated it under the rule of law attribute.

42. Goertz and Starr, ‘Introduction’, 2.43. Cf. Goertz and Starr, ‘Introduction’, 17; Ragin, ‘Fuzzy-Set Analysis of Necessary

Conditions’, 179.44. In fact, as Elman, ‘Explanatory Typologies’, fn. 15 stresses typological theory is basi-

cally a simpler version of Ragin’s QCA; the very premise of which is the notions ofnecessary and sufficient causation.

45. Ragin, Redesigning Social Inquiry.46. Ragin, ‘Set Relations in Social Research’, 297. When calibrating the original scores

into fuzzy set memberships, we have used what Ragin (Redesigning Social Inquiry)terms the ‘direct method of calibration’ using the scores of 10, 5.5, and 1 to designatefull membership, the crossover point (maximum ambiguity), and full non-member-ship, respectively, for all conditions and the outcome. This choice makes sense boththeoretically and empirically due to the way our definitions and hypotheses are formu-lated and the way the BTI dataset is constructed.

47. To mirror the BTI operationalization as closely as possible when using the FH, the elec-toral attribute is covered by the components Electoral Process and Political Pluralismand Participation. The political liberties attribute is covered by the componentsFreedom of Expression and Belief and Associational and Organizational Rights.Finally, the rule of law attribute is covered by the component Rule of Law. Regardingthresholds, matters are a bit more complicated than in the case of the BTI as we do nothave any linguistic qualifiers to rely on. We have recalibrated the scores from 0–100because the different components are not measured using the same range. Onceagain, we have employed a minimum score procedure to aggregate them whenneeded. When calibrating the original scores into fuzzy set memberships, againemploying the direct method of calibration, the scores of 100, 60, and 0 were used todesignate full membership, the crossover point (maximum ambiguity), and full non-membership, respectively.

48. See for example Moreno et al., ‘A Failed Index’; Patrick, ‘“Failed” States and GlobalSecurity’.

49. To calibrate the FSI data into fuzzy set scores (using the direct method of calibration),we used the scores of 30, 90, and 115, to designate full membership (delineatingcountries in FSI’s sustainable category), the crossover point (delineating countriesin the FSI categories of alert and warning), and full non-membership (approximatelythe worst score achieved by any country), respectively.

50. Mahoney, ‘The Elaboration Model and Necessary Causes’, 285. To quote: ‘If Z ispresent, X will always also be present. In this sense X is a redundant necessary cause.And if Z is absent, we already know that a necessary cause is missing, and thus theoutcome will be absent. The issue of whether X is present or absent seems irrelevant’.

51. Kitschelt, ‘Accounting for Postcommunist Regime Diversity’.52. We have also tried to combine this indicator with BTI’s education indicator into a

modernization index. The results based on this measure are practically identical tothose derived only using the wealth indicator.

Democratization 19

Page 20: Stateness first?

53. To the extent that modernization is accepted as necessary for social rights, this wouldimply the establishment of a causal chain in which (the deep factor of) stateness is anecessary condition for (the more proximate factor of) modernization which is thennecessary for social rights. Modernization could thus be construed as the ‘interveningvariable’ through which stateness makes for social rights (Mahoney, ‘The ElaborationModel and Necessary Causes’, 283). The same could of course be the case for otherintervening factors.

54. Cf. Kraxberger, ‘Failed States’.55. For example, Krasner, ‘The Case for Shared Sovereignty’.56. For example, Herbst, ‘Let Them Fail.57. Paradoxically, ‘a potentially effective treatment of stateness problems – coercion and

repression – is not available to democrats. As it happens, however, they are those whoneed it the most’ (Elkins and Sides, Seeking Stateness, 1).

Notes on contributorsJørgen Møller has a PhD from the European University Institute in Florence, Italy (2007)and is currently Associate Professor at the Department of Political Science, Aarhus Univer-sity, Denmark. His research interests include conceptualization of democracy, post-commu-nist political change, patterns of state formation, and qualitative methodology.

Svend-Erik Skaaning has a PhD from the Aarhus University (2007) and is currently Associ-ate Professor at the Department of Political Science, Aarhus University, Denmark. Hisresearch interests include conceptualization of democracy, conceptualization of civil liber-ties, political change in developing countries, patterns of state formation, and qualitativemethodology.

BibliographyBack, Hanna, and Axel Hadenius. ‘Democracy and State Capacity: Exploring a J-Shaped

Relationship’. Governance 21, no. 1 (2008): 1–24.Beetham, David. Democracy and Human Rights. Cambridge: Polity Press, 1999.Bowman, Kirk, Fabrice Lehoucq, and James Mahoney. ‘Measuring Political Democracy:

Case Expertise, Data Adequacy, and Central America’. Comparative Political Studies38, no. 8 (2005): 939–70.

Bratton, Michael, and Eric C. Chang. ‘State Building and Democratization in Sub-SaharanAfrica: Forwards, Backwards, or Together?’. Comparative Political Studies 39, no. 9(2006): 1059–83.

Burnell, Peter, and Peter Calvert. ‘The Resilience of Democracy: An Introduction’.Democratization 9, no. 1 (1999): 1–32.

Collier, David, and James E. Mahon. ‘Conceptual “Stretching” Revisited: AdaptingCategories in Comparative Analysis’. American Political Science Review 87, no. 3(1993): 845–55.

Collier, David, and Steven Levitsky. ‘Democracy with Adjectives: Conceptual Innovation inComparative Research’. World Politics 49, no. 3 (1997): 430–51.

Dahl, Robert A. Polyarchy: Participation and Opposition. New Haven, CT: Yale UniversityPress, 1971.

Dahl, Robert A. Dilemmas of Pluralist Democracy: Autonomy vs. Control. New Haven, CT:Yale University Press, 1982.

Dahl, Robert A. A Preface to Economic Democracy. Cambridge: Polity Press, 1985.Dahl, Robert A. Democracy and Its Critics. New Haven, CT: Yale University Press, 1989.

20 J. Møller and S.-E. Skaaning

Page 21: Stateness first?

Diamond, Larry. Developing Democracy: Toward Consolidation. Baltimore, MD: TheJohns Hopkins University Press, 1999.

Elkins, Zachary, and John Sides. Seeking Stateness. Manuscript, 2008. http://home.gwu.edu/~jsides/stateness.pdf (accessed January 29, 2010).

Elman, Colin. ‘Explanatory Typologies in Qualitative Studies of International Politics’.International Organization 59, no. 2 (2005): 293–326.

Evans, Geoffrey. ‘The Eclipse of the State? Reflections on Stateness in an Era ofGlobalization’. World Politics 50, no. 1 (1997): 62–87.

Freedom House. Freedom in the World, 2010. http://www.freedomhouse.org/template.cfm?page=15 (accessed January 29, 2010).

Fukuyama, Francis. ‘The Imperative of State-Building’. Journal of Democracy 15, no. 2(2004): 17–31.

Fukuyama, Francis. ‘“Stateness” First’. Journal of Democracy 16, no. 1 (2005): 84–8.Fukuyama, Francis. ‘Liberalism versus State-Building’. Journal of Democracy 18, no. 3

(2007): 10–13.Goertz, Gary. Social Science Concepts: A User’s Guide. Princeton, NJ: Princeton University

Press, 2006.Goertz, Gary, and Harvey Starr. ‘Introduction: Necessary Condition Logics, Research

Design, and Theory’, in Necessary Conditions: Theory, Methodology, andApplications, ed. Gary Goertz and Harvey Starr, 1–24. Lanham, MD: Rowman &Littlefield Publishers, 2003.

Grzymala-Busse, Anna, and Paulina Jones Luong. ‘Reconceptualizing the Post-CommunistState’. Politics and Society 30, no. 4 (2002): 529–54.

Hadenius, Axel, ed. Democracy’s Victory and Crisis. Cambridge: Cambridge UniversityPress, 1997.

Herbst, Jeffrey. ‘Let Them Fail: State Failure in Theory and Practice’, in When States Fail:Causes and Consequences, ed. Robert E. Rotberg, 302–18. Princeton, NJ: PrincetonUniversity Press, 2004.

Kaufman, Daniel, Aart Kray, and Massimo Mastruzzi. Governance Matters VIII: Aggregateand Individual Governance Indicators, 1996–2008. World Bank Policy ResearchWorking Paper No. 4978. Washington, DC: World Bank, 2009.

Kitschelt, Herbert. ‘Accounting for Postcommunist Regime Diversity: What Counts as aGood Cause?’, in Capitalism and Democracy in Central and Eastern Europe:Assessing the Legacy of Communist Rule, ed. Grzegorz Ekiert and StephenE. Hanson, 49–86. Cambridge: Cambridge University Press, 2003.

Krasner, Stephen. ‘The Case for Shared Sovereignty’. Journal of Democracy 16, no. 1(2005): 69–83.

Kraxberger, Brennan. ‘Failed States: Contemporary Obstacles to Democratic Diffusion’.Third World Quarterly 28, no. 6 (2007): 1055–71.

Levitsky, Steven, and Lucan A. Way. ‘The Rise of Competitive Authoritarianism’. Journalof Democracy 13, no. 1 (2002): 51–65.

Linz, Juan. ‘Democracy Today: An Agenda for Students of Democracy’. ScandinavianPolitical Studies 20, no. 2 (1997): 115–34.

Linz, Juan, and Alfred Stepan. Problems of Democratic Transition and Consolidation:Southern Europe, South America, and Post-Communist Europe. Baltimore, MD: JohnHopkins University Press, 1996.

Mahoney, James. ‘The Elaboration Model and Necessary Causes’, in Explaining War andPeace: Case Studies and Necessary Condition Counterfactuals, ed. Gary Goertz andJack Levy, 281–306. Abingdon: Routledge, 2007.

Marshall, T.H. Citizenship and Social Class. London: Pluto, 1996[1949].Moreno, Luis A., Pierre Schori, Eduardo Ferrero, Flavio D. Espinal, Andrews S. Natsios,

and Louis W. Goodman. ‘A Failed Index’. Foreign Policy 150 (2005): 4–12.

Democratization 21

Page 22: Stateness first?

Møller, Jørgen, and Svend-Erik Skaaning. ‘Beyond the Radial Delusion: Conceptualizingand Measuring Democracy and Non-Democracy’. International Political ScienceReview 31, no. 3 (2010): 261–83.

Møller, Jørgen, and Svend-Erik Skaaning. ‘Marshall Revisited: The Sequence of CitizenshipRights in the Twenty-First Century’. Government and Opposition 54, no. 4 (2010):457–83.

Munck, Gerardo, and Jay Verkuilen. ‘Conceptualizing and Measuring Democracy’.Comparative Political Studies 35, no. 1 (2002): 5–34.

Nodia, Ghia. ‘How Different Are Postcommunist Transitions?’. Journal of Democracy 7,no. 4 (1996): 15–29.

O’Donnell, Guillermo. ‘Polyarchies and the (Un)rule of Law in Latin America: A PartialConclusion’, in The (Un)rule of Law and the Underprivileged in Latin America, ed.Juan E. Mendez, Guillermo O’Donnell, and Paulo Sergio Pinherio, 303–37. NotreDame: Notre Dame University Press, 1999.

O’Donnell, Guillermo. ‘Democracy, Law, and Comparative Politics’. Studies inComparative International Development 36, no. 1 (2001): 7–36.

O’Donnell, Guillermo. ‘Human Development, Human Rights, and Democracy’, in The Qualityof Democracy: Theory and Applications, ed. Guillermo O’Donnell, Jorge V. Cullell, andOsvaldo M. Iazzetta, 9–92. Notre Dame: Notre Dame University Press, 2004.

O’Donnell, Guillermo. Dissonances: Democratic Critiques of Democracy. Indiana:University of Notre Dame Press, 2007.

Patrick, Stewart. ‘“Failed” States and Global Security: Empirical Questions and PolicyDilemmas’. International Studies Review 9, no. 4 (2007): 644–62.

Ragin, Charles. ‘Fuzzy-Set Analysis of Necessary Conditions’, in Necessary Conditions:Theory, Methodology, and Applications, ed. Gary Goertz and Harvey Starr, 179–96.Lanham, MD: Rowman & Littlefield Publishers, 2003.

Ragin, Charles. ‘Set Relations in Social Research: Evaluating Their Consistency andCoverage’. Political Analysis 14, no. 3 (2006): 291–310.

Ragin, Charles. Redesigning Social Inquiry: Fuzzy Sets and Beyond. Chicago, IL:University of Chicago Press, 2008.

Rose, Richard, and Doh Chull Shin. ‘Democratization Backwards: The Problem ofThird-Wave Democracies’. British Journal of Political Science 31, no. 2 (2001):331–54.

Sartori, Giovanni. ‘Concept Misformation in Comparative Politics’. American PoliticalScience Review 64, no. 4 (1970): 1033–53.

Sartori, Giovanni. ‘Guidelines for Concept Analysis’, in Social Science Concepts: ASystematic Analysis, ed. Giovanni Sartori, 15–85. London: Sage, 1984.

Schumpeter, Joseph A. Capitalism, Socialism and Democracy. London: Unwin UniversityBooks, 1974[1943].

Skaaning, Svend-Erik. ‘Measuring Civil Liberty: An Assessment of Standards-Based DataSets’. Revista de Ciencia Polıtica 29, no. 3 (2009): 721–40.

Skaaning, Svend-Erik. ‘Measuring the Rule of Law’. Political Research Quarterly 63, no. 2(2010): 449–60.

Sørensen, Georg. Democracy and Democratization. Boulder, CO: Westview Press, 2008.Tilly, Charles. Coercion, Capital and European States, Ad 990–1992. Oxford: Blackwell

Publishers, 1992.Tilly, Charles. Democracy. Cambridge: Cambridge University Press, 2007.Van Creveld, Martin. The Rise and Decline of the State. Cambridge: Cambridge University

Press, 1999.Weber, Eugen. Peasants into Frenchmen: The Modernization of Rural France. Stranford,

CA: Stanford University Press, 1976.

22 J. Møller and S.-E. Skaaning

Page 23: Stateness first?

Appendix 1: Describing the BTI indicators

Table A1. BTI’s subcomponents: links between scores and subcategory conditions.

Indicator Question Clarification

1.1 Monopoly onuse of force

To what extent does the state’s monopoly on the use offorce cover the entire territory?

1.2 Citizenshipagreement

To what extent do all relevant groups in society agree aboutcitizenship and accept the nation-state as legitimate?

The question seeks to assess the extent to which (1) majorgroups in society accept and support the official/dominantconcept of the nation-state and (2) ruling groups de jure or defacto exclude ethnic, religious or cultural minorities frompolitical citizenship (membership in the political nation)

2.1 Free elections To what extent are rulers determined by general, free andfair elections?

2.2 Democrats rule To what extent do democratically elected rulers have theeffective power to govern, or to what extent are thereveto powers and political enclaves?

The elected rulers to be assessed here are primarily thepresident and/or parliament or the head of government whois empowered by the parliament. Veto powers can be themilitary, the clergy, landowners, business elites, etc., withthe power to defect partially from democratic procedureswithout questioning the system as such.

2.3 Association/assembly rights

To what extent can independent political and/or civicgroups associate and assemble freely?

This question refers to the degree of freedom to organizepolitically that is needed to influence political decisionmaking processes ‘from the bottom up’. It also includesgroups that mobilize along ethnic and similar cleavages.

2.4 Freedom ofexpression

To what extent can citizens, organizations and the massmedia express opinions freely?

Apart from evaluating to what extent freedom of opinion andthe press are generally guaranteed, please also considerwhether the structure of the mass media system provides fora plurality of opinions

3.1 Separation ofpowers

To what extent is there a working separation of powers(checks and balances)?

This question refers to the basic configuration and operation ofthe separation of powers (institutional differentiation,division of labour according to functions and, mostsignificantly, checks and balances). It does include thesubjection of state power to the law

(Continued)

Dem

ocratization23

Page 24: Stateness first?

Table A1. Continued.

Indicator Question Clarification

3.2 Independentjudiciary

To what extent does an independent judiciary exist? An independent judiciary refers first and foremost to how farthe courts can interpret and review norms and pursue theirown reasoning free from the influence of rulers or powerfulgroups and individuals. This requires a differentiatedorganization of the legal system, including legal education,jurisprudence, regulated appointment of the judiciary,rational proceedings, professionalism, channels of appealand court administration.

3.3 Abuse of officeprosecuted

To what extent are there legal or political penalties forofficeholders who abuse their positions?

This question addresses how the state and society hold publicservants and politicians accountable and whether conflicts ofinterest are sanctioned.

3.4 Civil rightsensured

To what extent are civil rights guaranteed and protected,and to what extent can citizens seek redress forviolations of these liberties?

The civil rights in question refer to the protection of personalliberty against state and non-state actors, including the rightto life and security of person; freedom from slavery; equalitybefore the law and due process under the rule of law;freedom of movement; access to justice, bans ondiscrimination and freedom of religion.

10.1 Social safetynets

To what extent do social safety nets exist to compensate forpoverty and other risks such as old age, illness,unemployment or disability?

Social safety nets may be organized by the state or by society(private welfare institutions), and comprise a variablewelfare mix. They are ultimately intended to ensureinclusion of almost everyone in economic life.

10.2 Equalopportunity

To what extent does equality of opportunity exist? The question of equal opportunity is directed at finding out towhat extent individuals have equal access to participation insociety regardless of their social background. Stateinitiatives – such as access to public services, especiallyeducation, and assistance mechanisms – play a central role

24J.

Møller

andS.-E

.Skaaning