Sartori’s typology

The main innovation of Sartori’s (Reference Sartori1976) typology was to combine a traditional focus on the format (i.e. number of parties) with a more competition-centred approach that took into consideration the ideological distance between the polar opposites of the party system (Enyedi and Casal Bértoa, Reference Enyedi and Casal Bértoa2020). Such a combination (see Figure 1) allowed him to distinguish seven different party system types, the most prominent being “predominant,” “two-party,” “moderate” and “polarized” pluralist systems (Sartori, Reference Sartori1976: chapter VI). A full discussion of Sartori’s typology is provided in the online appendix section A.

Figure 1.

Sartori’s “simplified model”. Source: Sartori (Reference Sartori1976: 292).

Sartori and his critics

Notwithstanding its application in multiple settings (e.g. Mair, Reference Mair1987; Heidar, Reference Heidar, Mair and Smith1990; Pappalardo, Reference Pappalardo1996; Casal Bértoa, (Reference Casal Bértoa2013): Quinn, Reference Quinn2013; Ignazi, Reference Ignazi2017; Lee and Casal Bértoa, Reference Lee, Casal Bértoa, Kasuya and Tan2025), Sartori’s typology has been the object of heavy criticism. Perhaps the most drastic was Pennings (Reference Pennings, Pennings and Lane1998), which even merited Evans’ (Evans, Reference Evans2002) ad hominem defence. While there have been certainly many others (Daalder, Reference Daalder, Daalder and Mair1983; Loenen, Reference Loenen1990; Wolinetz, Reference Wolinetz, Katz and Crotty2006), we confine ourselves here to the most important: the hollowing out of the polarized pluralist category, caused by “the demise of the traditional anti-system [e.g. Communist, Fascist] party” (Mair, Reference Mair, Katz and Crotty2006: 65). The implications of this are mainly four.

First, the main consequence of the abovementioned hollowing is, clearly, the overcrowding of the moderate pluralist type. Secondly, while “[t]he Sartori model assumes a close link between fragmentation and polarization: the higher the number of parties, the higher the ideological distance in the system” (Enyedi and Casal Bértoa, Reference Enyedi, Casal Bértoa, Carter, Keith, Sindre and Vasilopoulou2023: 35), this does not, in fact, seem to be the case.Footnote ⁴ Thirdly, such conflation of party systems mostly into one single type makes the typology less appropriate to the task of identifying dynamics and making distinctions within the current world. In this context, and already implicit in Sartori’s work, it is clear that not all systems included in the “moderate pluralist” category behave the same: namely, their mechanics – to use Sartori’s expression – are very different. Finally, and most importantly from the empirical point of view, such an accumulation makes party system change, understood as change “from one class or type […] into another” (Mair, Reference Mair1997: 51; see also Casal Bértoa, Reference Casal Bértoa2013), almost impossible or extremely rare (for a full discussion of critiques, see Section B in the online appendix).

Democratic party systems in Asia: new dataset, classic operationalization

In order to test how well Sartori’s and other alternative typologies work outside Europe, we make use of a new dataset that includes 24 different party systems in 17 Asian democracies between the conclusion of the Second World War and the end of 2020 (Casal Bértoa and Lee, Reference Casal Bértoa and Lee2021). In Appendix Section C, we provide the details of our new dataset and include a comprehensive explanation of how we objectively operationalize all the variables used in our analysis.

Sartori goes to Asia: challenges and pitfalls

Figure 2 shows the relationship between (parliamentary) fragmentation and (political) polarization in all 24 democratic party systems in Asia. Using Sartori’s typology one can distinguish between (1) predominant party systems like Bangladesh I and Myanmar II; (2) two-party systems like Mongolia and Bhutan; (3) moderate pluralist party systems like East Timor, Japan, Pakistan I, II and III, Sri Lanka I and Taiwan; and (4) polarized party systems like Indonesia, Kyrgyzstan, Malaysia II, Nepal II and Thailand.Footnote ⁵

Figure 2.

(Parliamentary) fragmentation and (political) polarization in 24 Asian democracies. Source: Casal Bértoa and Lee (Reference Casal Bértoa and Lee2021)

One third of the Asian party systems, however, do not seem to fit easily into Sartori’s typology. Thus, one can find very polarized party systems like Myanmar I and Nepal II, non-polarized extreme party systems (e.g. Philippines), and, especially, a rather high number of limited but polarized party systems (e.g. Bangladesh II, India, Malaysia I, South Korea and Sri Lanka II).

Even if we were to (1) use Sartori’s subjective rules of counting, allowing us to classify Japan, Malaysia I and Myanmar I as predominant party systems,Footnote ⁶ and (2) regard the Indian and South Korean party systems as non-polarizedFootnote ⁷; a quick glance at Figure 2 shows the lack of statistical correlation between fragmentation and polarization in Asia (.186).

Secondly, Figure 2 also shows the clear accumulation of Asian party systems around the moderate pluralist type, which includes half of them: namely, the seven signalled above (East Timor, Japan, Sri Lanka I, Taiwan, and three in Pakistan), plus all those that, having an ENPP between 2.5 and 3.9, are highly polarized (e.g. Bangladesh II, Malaysia I, Sri Lanka II) or moderately so (e.g. India, South Korea).

Thirdly, when looking at the countries classified as polarized pluralist, we find that there is only one example of what Sartori would call “anti-systemic parties,” namely, the Malaysian Islamic Party (PAS). In Nepal, there are traditionally Communist parties like the Unified Marxist-Leninist (CPN/UML) or the Maoist Centre (CMP/MC), but they are at the centre of the party system, making it to every single coalition since re-democratization in the late 2000s. As a result, like in Europe (Mair, Reference Mair, Katz and Crotty2006), none of the party systems classified above as “polarized pluralist” really fits the mechanics, even if they fit the format, of “centrifugal” politics, at least not in the Sartorian way of having two or more irresponsible opposition parties located at the extremes of the political spectrum that oblige the centre to rule perpetually.

Fourthly, already implicit in Sartori’s work, it is important to note that not all moderate pluralist party systems behave the same. Thus, some, like Pakistan III, have more the logic of a (German-style) “two-and-a-half” party system, others, like Indonesia and Nepal II, behave more like the Dutch or Finnish party systems where three different party families (socialist, liberal and conservative) alternate in government, still others are characterized, as in Denmark and Hungary, by the opposition of two ideologically inimical blocs of parties.

Finally, there is the everlasting problem of party system change. Considering, like Sartori (Reference Sartori1976) and Mair (Reference Mair, Katz and Crotty2006), that party system change takes place when there is a change in type, we can say that with the exception of Bangladesh I, Bhutan, Nepal I, Pakistan I, Myanmar II (all too short-lived to experience any systemic change), Kyrgyzstan, Indonesia, Malaysia II, Pakistan II and Sri Lanka II, all other Asian party systems experienced at least one change. However, as we will see later, this is problematic for two reasons.

On the one hand, it clearly underestimates systemic change. For example, any observer of Indonesian politics will signal an important change in the party system after the 2014 presidential/legislative elections caused by the institutionalization of political parties and the reform of the electoral system five years earlier (Trihartono and Patriadi, Reference Trihartono and Patriadi2016; Okthariza, Reference Okthariza, Teehankee and Echle2023).

On the other hand, it overestimates party system change. Four examples will suffice to illustrate this point. While the Taiwanese party system has remained a two-party system, pitting DPP on the left against KMT on the right, since the very beginning (Lee, Reference Lee2024), an application of the indicators mentioned above would suggest a change from two-partyism towards moderate pluralism in 1995, with a return to two-partyism in 2012 after a period of KMT’s dominance between 2008 and 2012. Similarly, the Philippine party system, characterized by extremely high fragmentation among unstable parties (Teekankee, Reference Teekankee2020), would be considered according to Sartorian rules to alternate some periods under the moderate pluralist type in 1992–2000, 2004–2006 and 2010–2018. Similarly, Bangladesh II, where two blocs of parties around AL and BNP alternate in power continuously between 1991 and 2006 (Rahman, Reference Rahman2019), would have changed from a moderate pluralist type to a two-party system in 2001. But perhaps the clearest example is East Timor. Having been characterized between independence in 2001 and 2017 by the opposition between the Revolutionary Front for an Independent East Timor (Freitlin) and the rest of the parties (Shoesmith, Reference Shoesmith2020), it would be considered to have left two-partyism to adopt a polarized pluralist type in 2007, before changing to a moderate pluralist type in 2012.

A new typology of Asian party systems

Recently, Casal Bértoa and Enyedi (Reference Casal Bértoa and Enyedi2021) have proposed a new typology where they distinguish eight types of party systems, considering the party configuration and likely extent of polarization, as well as the number of poles. These eight types can be categorized into five: unipolar, bipolar, tripolar, multipolar and centre-based. A full discussion of these different types can be found in the online appendix (Section D).

Figure 3 shows a typology of Asian party systems according to their level of institutionalization, measured using Casal Bértoa and Enyedi’s (Reference Casal Bértoa and Enyedi2021) index of closure.Footnote ⁸ While Table A2 in the Appendix contains a summary of the different structures of inter-party competition and party system types observed in each of the countries analysed per period, Figure 3 classifies each party system according to the type that dominated.

Figure 3.

Party system type according to the level of institutionalization (i.e. closure). Source: Casal Bértoa And Lee (Reference Casal Bértoa and Lee2021)

As follows from the figure above, the only party system type to be absent in the region – in contrast to Europe – is the centre-based one, which would be equivalent to Sartori’s (Reference Sartori1976) extreme polarized multi-party system. Among the rest, bipolar types have been the most popular (see Table A2): two-bloc (14 times, including twice in Pakistan III, Sri Lanka I/II, as well as Thailand),Footnote ⁹ two-party (6), and two-plus-one (5).Footnote ¹⁰

In clear contrast to Europe, where no dominant party system has ever existed (see Casal Bértoa and Enyedi, Reference Casal Bértoa and Enyedi2021: 122), eight party systems in Asia experienced a dominant pattern of inter-party competition. Most of them were formed immediately after independence (e.g. Bangladesh I, India, Malaysia I/II, Myanmar I). Others were formed after the merger of two main parties in the mid-1950s (e.g. the Liberal and the Democratic parties in Japan), the complete collapse of the once powerful democratic opposition party (e.g. the Democratic Party in Mongolia), or the country’s transition to democracy (Myanmar II in 2016). The majority ended with the collapse of democracy, usually after a military coup (e.g. Bangladesh I, Myanmar I),Footnote ¹¹ a Communist insurgency (e.g. Malaysia I), or simply due to the electoral decline of the dominant party (e.g. the Indian National Congress in 1989,Footnote ¹² UNMO in 2018). In Japan, it was the 1994 reform that introduced a new – more proportional – electoral system (i.e. SNTV was replaced by MMP).

Both the Kyrgyz and Philippine party systems have been characterized by multipolarity, a pattern that was also visible in Thailand between 1992 and the beginning of the 21^st century. A tripolar structure of competition appears only three times: in Indonesia between 2004 and 2013, in Nepal II after 2008, and, more recently, in Pakistan III after 2018. Grand coalitions are rare, although Mongolia has tried this formula a couple of times (2004–2005, 2008–2011), with newly democratized Indonesia being the only other country to have briefly (1999–2003) experienced such a “broad” pattern.

Party system change, understood as taking place when adopting a new party system type (Casal Bértoa, (Reference Casal Bértoa2023)), took place in more than half of the party systems here studied (for more details, See section E in the online appendix). This clearly contrasts with Europe, where only 24 party systems did not experience systemic change (see Casal Bértoa and Enyedi, Reference Casal Bértoa and Enyedi2021: 123; Casal Bértoa, Reference Casal Bértoa2024). Some, like Bangladesh II, Kyrgyzstan, Myanmar I, Nepal II, Pakistan II, the Philippines, Sri Lanka I/II, and Taiwan experienced no change in their structure of inter-party competition whatsoever.Footnote ¹³ Not surprisingly, the same can be said of those short-lived party systems where free party competition did not manage to last even a decade, either because democracy collapsed rather quickly (e.g. Bangladesh I, Myanmar II, Nepal I, and Pakistan I) or because its democratic transition only happened very recently (e.g. Bhutan).

Focusing only on the type of party system that has been predominant in each country, we can distinguish five different clusters: dominant Bangladesh I, India, Japan, Malaysia I and II as well as Myanmar I and II; two-party Bhutan, Nepal I, Pakistan I, South Korea and Taiwan, two-bloc Bangladesh II, East Timor, Mongolia, Pakistan II, Sri Lanka I and II; two-plus-one Pakistan III; tripolar Indonesia and Nepal II; and multipolar Kyrgyzstan, Philippines and Thailand. In addition, in Appendix Table A3, we present the comparison between Asian and European party system types.