This is a guest post by Grigorii V. Golosov. It is based on the article “The Five Shades of Grey: Party Systems and Authoritarian Institutions in Post-Soviet Central Asian States”, in Central Asian Survey, doi:10.1080/02634937.2018.150044.
As of 2017, there was no single instance of a state that could be qualified as an electoral democracy in post-Soviet Central Asia. Meanwhile, all these states do conduct elections on a fairly regular basis, and they have developed party systems that display a remarkable cross-national variation. While the consequences of authoritarian institutions have been already subjected to a rather close scrutiny in scholarly literature, relatively little is known about the reasons why such institutions take their specific shape. This problem is particularly salient with respect to the most fundamental property of party systems, fragmentation, conventionally defined as the number of important parties in the system. While it is certainly true that the general tendency of authoritarian regimes is to create dominant party systems in which a pro-government party takes a lion’s share of votes and legislative seats, there are well-known instances of long-standing yet very fragmented authoritarian party systems, such as in Morocco. The reasons for this situation have not been clarified in the literature. The purpose of this study is to solve this puzzle by means of a systematic comparison of five authoritarian party systems that, while being quite similar in their origins, belonging to the same region of Post-Soviet Central Asia and thus sharing many similarities of context, do display a significant variation on the main parameter of interest, fragmentation.
In this study, I use a conventional measure of fragmentation, the effective number of parties, in the mathematical formulation developed by Golosov: where pi and p1 stand for the fractional shares of seats received by the i-th and the largest parties, respectively. Table 1 provides a brief summary of parliamentary elections, presented as the observed levels of party system fragmentation, in Central Asian states throughout the whole post-Soviet period. Since the electoral cycles of individual countries do not concur, each of the cells of the table reports the year of elections (or the year of the first round of elections, if different from that of the second round), and the effective number of parliamentary parties in parentheses.
Abbreviations: NPS – no party structure in the assembly; SPS – single-party system. Sources: Inter-Parliamentary Union 2017; Nohlen, Grotz and Hartmann 2001.
I hypothesize that the levels of party system fragmentation in autocracies are contingent upon the scope of presidential powers. This is not to say that formal constitutional provisions, shallow as they are in authoritarian political contexts, are matters of primary concern for the autocrats. Of course, they are primarily concerned with preserving and expanding their hold over the polity. Yet if they choose to follow the rules of the game as established in the constitution, then it is rational for them to facilitate the development of secondary political structures, such as party systems, in a way that is optimal for the consolidation of authoritarianism within the existing constitutional order. In other words, the autocrats are well motivated to seek congruency between the scope of their constitutional powers and party system properties. In fact, there is no reason to assume that in this respect, the autocrats are fundamentally different from democratic presidents. What differentiates the two is not motivation but rather capacity: the autocrats are much better equipped to shape party systems at their will.
For an institutionally strong president, the party composition of the assembly is not very consequential. Given that the president makes crucial political and administrative decisions without any parliamentary involvement and often possesses significant legislative powers, political fragmentation in the assembly can be affordable for the presidency. In democracies, the limits of such affordability are set by the president’s ultimate inability to pursue her legislative agenda in a complete defiance of the parliament, which explains why the coexistence of a strong presidency with a fragmented assembly is often problematic. In autocracies, a strong president is less constrained by such considerations. At the same time, parliamentary fragmentation offers several tangible rewards to an institutionally strong president. First, the presence of many parties in the assembly greatly increases the symbolic value of the parliament by providing a visible proof to the autocrat’s claim that the regime is democratic in its nature. Indeed, one of the primary goals of authoritarian institution building is to project an image of democracy without threatening the political survival of the autocrat. Second, the survival of autocrats in power largely depends on their ability to neutralize the actual or potential opponents by co-opting them into the institutional structure of the regime. This explains why cooptation is often viewed as the principal rationale for the very existence of authoritarian institutions. Yet for co-optation to occur, a wide range of elite groups – including those that do not belong to the autocrat’s ‘inner circle’ – should gain access to representation, however shallow, and to the spoils associated with it. Third, by allowing such groups to participate in elections under their own party labels, the regime makes use of their political machines, which is important in authoritarian electoral contexts that are often dominated by clientelism.
Taking this into account, consider a situation when the formal powers of the presidency are intermediate, neither very strong nor very weak. In this situation, the autocratic president still can afford a multi-party composition of the assembly. However, the payoff of this strategy is smaller than in the conditions outlined above. For example, if the president is involved in the formation of the government yet the parliament also has significant appointment / government censure powers, which is often the case when the constitutional powers of the presidency are within an intermediate range, then manufacturing a multiparty majority poses a problem for the president without significantly reducing the political capacity of the legislature. In this situation, it is in the best interest of the president to seek and obtain a single-party legislative majority. A fragmented legislature is not necessarily dangerous for an autocratic president with moderate formal powers, but it is certainly more difficult to handle. This creates a strong incentive for seeking a pro-presidential party majority in the legislature. In the presence of strong motivation towards single-party control over the assembly, those factors that have been listed above as facilitating the coexistence between institutionally strong presidents and fragmented assemblies become less consequential. Thus the main hypothesis of this study can be formulated as follows. There is a curvilinear pattern of association between the constitutional powers of the presidency and the effective number of parties in electoral authoritarian regimes. While institutionally weak and institutionally strong presidencies are associated with fragmented party systems, the autocratic presidents with moderate yet sizeable formal powers seek and obtain single-party majorities in their assemblies, which leads to low levels of party system fragmentation.
In order to test this hypothesis, I used the Prespow scores as reported at this site with some minor amendments. Table 2 presents the scores as used in this study. The structure of presentation is the same as in Table 1.
Sources: Inter-Parliamentary Union 2017; Presidential Power 2017.
With the data on the powers of the presidency and the effective number of parties at hand, I proceeded with my statistical analysis. The overall number of observations in the analysis is 16, which corresponds to the number of party-structured elections held in Central Asian states throughout the period under observation, as defined above and reported in Tables 1 and 2. This is explained in detail in the previous section of this paper. Given that the relationship between the two variables of interest is expected to be curvilinear, I used a polynomial regression model. The scatter plot is provided in Figure 1 (note that two data points for Kazakhstan, 2012 and 2016, nearly coincide in space). As is evident from the figure, the trend line takes the form of a U-shaped parabola, which supports my expectations regarding the relationship between the constitutional powers of the presidency and the effective number of parties.
The observed relationship is very strong, with R-squared = 0.67, and statistically significant at 0.001. The regression equation is thus: where ENP stands for the effective number of parties.
The strength of the observed association between the scope of presidential powers and the effective number of parties in Central Asian states suggests that the presidents can direct political processes – in this case, the process of party system development – at their will. Here lies the main difference between the dynamics of party system development in democracies, where it is largely determined by societal cleavages and other factors beyond the control of the acting rulers, and in autocracies, where it is largely manipulated by the authorities. Of course, a wider cross-national inquiry is needed to fully substantiate this vision of institutional dynamics under electoral authoritarianism, and it might well be that the dynamics observed in Central Asia deviates from the general tendency.
Sources: see sources to Tables 1 and 2.