¹ In Frost, Robert, Mountain Interval, New York, Henry Holt, 1921, p. 9 Google Scholar.

² Marshall, T. H., Citizenship and Social Class, London, Pluto, 1996 [1949], p. 10 Google Scholar.

³ See Mann, Michael, ‘Ruling Class Strategies and Citizenship’, in Bulmer, Martin and Reeve, Anthony (eds), Citizenship Today, London, UCL Press, 1996, ch. 7Google Scholar.

⁴ See Reinhard Bendix, Nation-Building and Citizenship, New York, John Wiley, 1964. Not all agree, though. Mann criticizes Marshall's model for being too Anglophile and evolutionary. See Mann, Ruling Class Strategies, p. 128. Likewise, Hirschman notes that ‘the Marshallian story – the progression from civil rights to mass participation in politics through universal suffrage to socioeconomic entitlements – proceeded in a far more leisurely and “orderly” manner in Great Britain than in the other major European countries, not to speak of the rest of the world’, Albert O. Hirschman, The Rhetoric of Reaction, Cambridge, MA, Harvard University Press, 1991, p. 128.

⁵ See, e.g. O'Donnell, Guillermo, ‘Democracy, Law, and Comparative Politics’, Studies in Comparative International Development, 36 (Spring 2001), pp. 7–36 CrossRefGoogle Scholar; Georg Sørensen, Democracy and Democratization, Boulder, CO, Westview Press, 2008; Larry Diamond, Developing Democracy: Toward Consolidation, Baltimore, Johns Hopkins University Press, 1999; Fareed Zakaria, The Future of Freedom: Illiberal Democracy at Home and Abroad, New York, Norton, 2003.

⁶ Guillermo O'Donnell, ‘Polyarchy and the (Un)Rule of Law in Latin America’, paper presented at the Meeting of the Latin American Studies Association, Chicago, September 1998, pp. 11–12.

⁷ O'Donnell, ‘Democracy, Law, and Comparative Politics’, p. 23.

⁸ See, e.g. Sørensen, Democracy and Democratization; Guillermo O'Donnell, Dissonances: Democratic Critiques of Democracy, Notre Dame, IN, University of Notre Dame Press, 2007.

⁹ Diamond, Developing Democracy, p. 57; O'Donnell, Dissonances, p. 121.

¹⁰ Oxhorn, Philip, ‘Social Inequality, Civil Society, and the Limits of Citizenship in Latin America’, in Eckstein, Susan and Wickham-Crowley, Timothy (eds), What Justice? Whose Justice? Fighting for Fairness in Latin America, Berkeley, University of California Press, 2003 Google Scholar.

¹¹ Levitsky, Steven and Way, Lucan, ‘The Rise of Competitive Authoritarianism’, Journal of Democracy, 13 (April), p. 61;Google Scholar Grzymala-Busse, Anna and Luong, Paulina Jones, ‘Reconceptualizing the Post-Communist State’, Politics and Society, 30 (December 2002), p. 536 CrossRefGoogle Scholar.

¹² See O'Donnell, Guillermo, ‘Horizontal Accountability in New Democracies’, Journal of Democracy, 9 (July 1998), pp. 117–18CrossRefGoogle Scholar; David Beetham, Democracy and Human Rights, Cambridge, Polity Press, 1999; Michael Saward, Democracy, Cambridge, Polity Press, 2003. Oxhorn basically makes the same point with explicit reference to Marshall: ‘The unprecedented ascendance of political democracy, with its concomitant guarantees of political rights, has shifted attention away from Marshall's original focus on the evolution of citizenship rights.’ Oxhorn, Social Inequality, p. 36.

¹³ Diamond, Developing Democracy, pp. 55–6.

¹⁴ Oxhorn, Social Inequality, p. 36.

¹⁵ Foweraker, Joe and Krznaric, Roman, ‘The Uneven Performance of the Democracies of the 3rd Wave: Electoral Politics and the Imperfect Rule of Law in Latin America’, Latin American Politics and Society, 44 (September 2002), p. 32 CrossRefGoogle Scholar.

¹⁶ O'Donnell, ‘Democracy, Law, and Comparative Politics’, p. 55.

¹⁷ Two sub-components (namely, personal social freedoms and equality of opportunity) of one of Freedom House's civil liberty attributes, namely, personal autonomy and individual rights, constitute partial exceptions to this observation. But unfortunately data for this (lowest) level of measurement is not publicly available.

¹⁸ Primarily by using objective indicators such as child mortality, life expectancy, literacy and GDP per capita.

¹⁹ Hence, we use the kind of typology that Bailey terms ‘Classification, then Identification’, meaning that first we create our conceptual typology and then order the empirical referents within it. Furthermore, we use a monothetic typology, meaning that the ‘possession of a unique set of features is both necessary and sufficient for identifying a specimen as belonging to a particular cell of the typology’. Bailey, Kenneth, ‘Monothetic and Polythetic Typologies and their Relationship to Conceptualization, Measurement and Scaling’, American Sociological Review, 38 (February 1973), pp. 21, 27 CrossRefGoogle Scholar.

²⁰ Oxhorn, Social Inequality, p. 37.

²¹ Marshall, Citizenship, p. 10.

²² To the extent that scores from more than one subcategory are used to capture an attribute, we consider them to be mutually constitutive. Consequently, rather than using the average, we employ a minimum score procedure to aggregate them, as recommended by Gary Goertz, Social Science Concepts: A User's Guide, Princeton, Princeton University Press, 2006. See also Bowman, Kirk, Lehoucq, Fabrice and Mahoney, James, ‘Measuring Political Democracy’, Comparative Political Studies, 38 (2005), pp. 939–70CrossRefGoogle Scholar.

²³ The respective positioning of the types could take many other shapes but that would not change the identity of the 27 types – only their position in the eye of the beholder.

²⁴ Bailey, Monothetic and Polythetic. The perfect simple order scale, also known as a Guttman scale, was first introduced in the context of statistical surveys. A perfect scale of this kind consists of a set of items that are ranked in order of difficulty from least extreme to most extreme position. For example, a person scoring a ‘6’ on a ten-item scale, will agree with items 1–6 and disagree with items 7, 8, 9, 10.

²⁵ This measure goes under the term ‘the coefficient of reproducibility’. The extent to which a typological ordering fits a perfect simple order scale is normally calculated using this coefficient which basically expressed the percentage of original values that can be reproduced by knowing the scale scores used to summarize them. We argue that this measure can be employed more generally to take stock of the empirical fit of a given typological theory, i.e. a particular hypothesis about the empirical distribution in a conceptual property space. This is so because the coefficient expresses the proportion of referents classified in accordance with the theoretical expectations.

²⁶ Anticipating the two robustness tests for 2007 carried out below, it is also this type that contains the majority of the outliers in these instances.

²⁷ Of course, this is a classical question of the glass being half full or half empty. The reason we have chosen the FH in the first place is that it is the only other widely employed dataset which makes for distinguishing between at least two of the three attributes.

²⁸ Regarding thresholds, matters are a bit more complicated than in the case of the BTI as we do not have any linguistic qualifiers to rely on, a lack of meaning attributed to the scores for which the FH has – in our opinion rightly – been criticized. See, e.g., Munck, Gerardo and Verkuilen, Jay, ‘Conceptualizing and Measuring Democracy’, Comparative Political Studies, 35 (2002), pp. 5–34 Google Scholar . Once again, we have employed a minimum score procedure to aggregate the sub-components when needed.

²⁹ There are some validity problems here as the Freedom House ratings cover more than the political and civil citizenship rights that Marshall had in mind. For instance, the attribute of ‘political rights’ includes aspects such as whether the government is accountable to the electorate between the elections and whether the government is free from pervasive corruption. However, these particular validity problems do not change the fact that generally speaking the Freedom House distinction comes close to that of Marshall.

³⁰ Zakaria, The Future of Freedom.

³¹ Notice, furthermore, that our results question the empirical underpinnings of the so-called claim about a gap between electoral and liberal democracy. This is so because it turns out that it is not countries having free and fair elections (no defects on political rights) in the absence of the civil liberties and rule of law that are currently predominant. Rather, it is countries exhibiting moderately defective elections in the absence of the said properties that is the most commonly encountered thinner type of democracy today.

³² Carothers, Thomas, ‘The Sequencing Fallacy’, Journal of Democracy, 18 (January 2007), pp. 12–27 CrossRefGoogle Scholar.

³³ See Krasner, Stephen, ‘The Case for Shared Sovereignty’, Journal of Democracy, 16 (January 2005), pp. 69–83 CrossRefGoogle Scholar.