Abstract
This article describes two opposing types of political personalization: centralizing and decentralizing personalization. The first implies the centralization of political power in the hands of a few leaders, while the latter indicates a diffusion of group power among its components: individual politicians. We start by proposing definitions of the types and subtypes of centralized and decentralized personalization and review the literature in search of evidence of their occurrence. We then demonstrate the usefulness of the proposed typology by examining personalization trends in various aspects of Israeli politics and conclude with a discussion of the challenges that personalization set for liberal democracies.
Introduction
Personal politics have always been part and parcel of politics, which is, after all, a product of interactions between individuals. Indeed, the reliance of political leadership on personal character has long been recognized (Farrell, 1971, cited in Poguntke and Webb, 2005b). However, in recent decades, several scholars have noted a trend of increasing personalization in democratic politics. Studies show a mutual process of growing centrality of the individual together with a decrease in the power of political groupings such as political parties, parliaments and cabinets (Blumler and Kavanagh, 1999; Mazzoleni and Schulz, 1999; McAllister, 2007; Poguntke and Webb, 2005a; Rahat and Sheafer, 2007; Wattenberg, 1991). Despite this growing interest in personalization, there remains a need for a conceptual and theoretical understanding of it and of its potential influence on democratic politics.
This article contributes to the study of personalization by pointing at two apparently opposite types of political personalization: centralizing and decentralizing personalization. Centralized personalization implies that power flows upwards from the group (e.g. political party, cabinet) to a single leader (e.g. party leader, prime minister, president). This is a common perception of personalization that results in the centralization of political power in the hands of a few leaders. An example might be the presidentialization of politics, a process in which the executive leader in parliamentary systems gains more autonomy relative to political parties and cabinets (Poguntke and Webb, 2005a). Decentralized personalization means that power flows downwards from the group to individual politicians who are not party or executive leaders (e.g. candidates, members of parliament, ministers). One example is the democratization of candidate selection, so that more candidates in democracies are selected by more inclusive selectorates in an explicitly more personalized competition (Hazan and Pennings, 2001). Note that decentralized personalization refers to power dispersion among the members of an elite group (in the example above, party representatives in parliament); it is not about power flowing down to rank-and-file members of the party.
Although centralized and decentralized personalization are more or less opposite processes, each does not necessarily come at the expense of the other and they can exist simultaneously. One result of these dual processes is that at the same time that leaders gain more personal power – so that they are no longer ‘first among equals’ but rather ‘first above equals’, they encounter more difficulties running the party they lead. This is because other individual members of the group, such as members of parliament, have also gained more personal power and can act independently, less restricted by their political party or its leader.
The study of personalization has not heretofore distinguished between centralized and decentralized personalization. Most studies focus on what we term centralized personalization. Decentralized personalization receives less attention in the literature, and when it does it is usually incidental. Therefore, the two main contributions of this article are first, the theoretical and empirical distinction between these two types of personalization, and second, new empirical evidence of decentralized personalization.
The first part of the article is devoted to presentation of a framework for the analysis of the different types of personalization. It includes definitions of the types and subtypes of centralized and decentralized personalization and reviews the literature in search of evidence of their occurrence. The second part examines personalization trends in various aspects of Israeli politics. A general trend of personalization, which has been identified empirically in previous studies, makes Israel a good case via which to demonstrate the two routes of personalization (Rahat and Sheafer, 2007; Shenhav and Sheafer, 2008). The article ends with a discussion of the challenges that these simultaneous processes of personalization set for democracies.
Centralized and decentralized personalization: Definitions and evidence
The framework presented here (Table 1) is based on Rahat and Sheafer’s (2007) typology of political personalization. They differentiate between institutional, media and behavioural personalization and propose some subtypes in the case of the latter two. Here we add a distinction between two types of personalization: centralized personalization and decentralized personalization.
Centralized and decentralized personalization: Types, sub-types and empirical evidence of their existence in established democracies.
‘Available evidence’ is the extent of empirical evidence of each type and sub-type of personalization. The available evidence can be evaluated as extensive, moderate, minimal or none (no available evidence). ‘Level of support’ is the extent to which the available evidence supports the claim for the occurrence of each type and sub-type of personalization. ‘Level of support’ can be categorized as support, mixed (supporting and challenging) or opposing evidence.
Institutional personalization
Institutional personalization implies ‘the adoption of rules, mechanisms, and institutions that put more emphasis on the individual politician and less on political groups and parties’ (Rahat and Sheafer, 2007: 207). These institutions grant more power to an individual (i.e. prime ministers, presidents, party leaders) in the case of centralized personalization and to individuals (i.e. ministers, legislators) in the case of decentralized personalization.
Centralized institutional personalization means that the leader’s (i.e. prime minister, president, party leader) personal position within an institution (i.e. cabinet or a party) is enhanced. In their study of the ‘presidentialization’ of politics in advanced democracies, Webb and Poguntke (2005: 377) identified a ‘shift in intra-executive power to benefit the head of the government’ and a shift of power within parties towards the party leader. This, they argue, implies ‘a shift from collective (or organizational) to individual power and accountability’ (op. cit., p. 352) (for more supportive evidence, see Bakvis and Wolinetz, 2005; Calise, 2005; Fiers and Krouwel, 2005; Hallin and Mancini, 2004; Lobo, 2005; Paloheimo, 2005; Pedersen and Knutsen, 2005). Party members have also increasingly been granted a central or even exclusive role in leadership selection over recent decades (Kenig, 2009a; Leduc, 2001). A similar trend is evident at the local government level, where more countries and municipalities have adopted direct mayoral elections (Magre and Bertrana, 2007). Direct elections of party leaders and of mayors imply personalization because they grant a personal popular (rather than partisan) mandate to those elected and because they render explicitly personal competition for these posts. Yet one should remember that many cases point to only a slight change towards strengthened institutional influence for the prime minister (O’Malley, 2007) and in other cases institutions remain largely unchanged (Bakvis and Wolinetz, 2005; van Biezen and Hopkin, 2005). The evidence available for this type of personalization is extensive but mixed.
Decentralized institutional personalization means that the personal position of individuals within an institution (a cabinet or party) is enhanced. The increasing weight of the personal vote in the electoral systems of some democracies (Austria, Belgium, The Netherlands and Sweden) is an example of such a trend (Karvonen, 2010). In these cases, candidates become more dependent on their own personal reputations, as opposed to the reputation of their parties (Shugart, 2001). This trend has also been identified within parties. In parallel with membership empowerment in leadership selection, more parties have given their members a central and even exclusive role in candidate selection (Bille, 2001; Caul-Kittilson and Scarrow, 2003; Scarrow et al., 2000). These trends imply personalization because – as in the case of party leaders and mayors -- they grant a personal popular mandate to those selected and elected, thus rendering competition for these posts more personal. The evidence obtainable from existing studies is moderate but does give support to the claim that personalization indeed occurs.
Media personalization
Media personalization refers to a change in the presentation of politics in the media, as expressed in a heightened focus on individual politicians and a diminished focus on parties, organizations and institutions (Blumler and Kavanagh, 1999; Mazzoleni and Schulz, 1999; Peri, 2004; Rahat and Sheafer, 2007; Reinemann and Wilke, 2001, 2007; van Aelst et al., 2012; Wattenberg, 1991). A further distinction can be made between political personalization in the unpaid media (i.e. media coverage of politics) and the paid media (i.e. political advertisement).
Unpaid media
Political personalization in the unpaid media refers to a change in the media coverage of politics in general, and of political campaigns in particular, manifested in an increase in journalists’ focus on the activities of individual politicians, at the expense of collective entities such as parties or of other political organizations and institutions.
Centralized unpaid media personalization means that journalists increasingly focus political coverage on a few leaders (presidents, prime ministers, party leaders) rather than on collectives such as cabinets and parties (van Aelst et al., 2012). There is a large body of evidence for a trend of centralized personalization in the unpaid media in democracies such as the UK (Langer, 2006, 2007) Japan (Krauss and Nyblade, 2005), Croatia (Grbeša, 2004), Israel (Rahat and Sheafer, 2007), France and Austria (McAllister, 2007), the U.S. (Wattenberg, 1998) and lately, to some extent, also Belgium (van Aelst, 2007).
However, not all scholars identify a clear cut trend of media centralized personalization. In The Netherlands, for example, Oegema and Kleinnijenhuis (2000) found that although party leaders play an important role in the news, political parties still dominate the text of the news items. In Germany, Reinemann and Wilke (2001, 2007) found no evidence of a process of media personalization, probably because focus on the Chancellor has always been very high. Analysis of Swedish news also did not support the existence of centralized personalization (Johansson, 2008). Evidence for this type of personalization is therefore extensive but mixed and to some extent contradictory.
Decentralized unpaid media personalization means an increased focus of the media on individual politicians who are not heads of the executive or party leaders. In contrast with centralized personalization, we do not have empirical evidence of decentralized personalization in the unpaid media. We know that the adoption of wider selectorates for candidate selection renders the media more important for politicians (Sheafer and Tzionit, 2006; Shenhav and Sheafer, 2008), yet we do not have evidence of a change over time in media coverage of individual candidates competing against each other.
Paid media
Political personalization in the paid media mainly refers to a change in the management of political campaigns. As part of the development of campaign strategies and tactics, campaign managers increasingly promote individual candidates at the expense of political parties and organizations (Bennett and Manheim, 2001; Brettschneider and Gabriel, 2002; Diamond and Bates, 1992; Holtz-Bacha, 2002, 2004; McAllister, 2007; Rahat and Sheafer, 2007; Swanson and Mancini, 1996).
Centralized personalization in the paid media implies that campaigns become more focused on leaders and less on their parties. Heffernan and Webb (2005) found evidence of this trend. They demonstrate that since 1960, General Election campaigns in Britain have become more candidate-centred, with parties offering leaders greater prominence in their election campaigns. Similarly, Boiney and Paletz (1991: 12), who conducted a review of political advertising literature in the US, concluded that [leaders’] ‘image appeals have become increasingly prominent, made either directly in the form of candidate qualities or conveyed indirectly using issues’ (for more supportive evidence, see Fiers and Krouwel, 2005; Toonen 1992). Yet other studies do not provide support for this trend. Holtz-Bacha (2004), for example, who conducted an empirical analysis of political spots on German television, found no overall trend of personalization, even for the large parties (see also, Lessinger and Holtz-Bacha, 2003). Similar negative results were found in Belgium (Bouckaert, 2009). There is, therefore, a moderate number of empirical analyses of this kind of personalization, and their findings are mixed.
Decentralized personalization in the paid media implies that campaigns become more focused on individual candidates (other than the party leader) and less on their parties. Empirical evidence of decentralized personalization in the paid media is less common in the political communication literature. However, an anecdotal illustration comes from Finland. Karvonen (2010) analysed all Finnish parliamentary campaign advertisements during the final week of the 1962, 1975, 1991 and 2007 elections. He divided the advertisements into three groups: those focusing on the party and not individual candidates; those focusing on several candidates but not on a complete list of candidates; and those including the name and candidate number or picture of an individual candidate. He found that over time ‘[t]he share of pure party ads has declined, and the share of individual ads has grown’ (ibid., p. 98). There is, therefore a minimal amount of empirical evidence for this type of personalization, but the evidence available supports the existence of decentralized personalization in the paid media.
Behavioural personalization of politicians
Behavioural personalization of politicians implies an increase in individualized political activities and a decline in collective partisan activity.
Centralized behavioural personalization of politicians implies that political leaders (e.g. prime ministers, presidents or party leaders) behave as individual actors, distancing themselves from the collectivities to which they belong (party and/or cabinet), or even ignoring them altogether. A few examples of centralized behaviour of leaders can be identified. In Italy, for instance, Prime Minister ‘Berlusconi felt strong enough to dismiss two of his most prominent ministers, an unprecedented move in the history of Italian government. As a consequence, he himself served as foreign minister for several months’ (Calise, 2005: 103). Calise argued that there was no doubt Berlusconi’s ascent to power had had an extraordinary personal impact, elevating personalization to its current level in Italy. In Japan, it was found that later prime ministers have been much more active than those of the 1970s, and in some cases have delivered more than twice the number of speeches and held twice the number of meetings in constituencies compared to earlier prime ministers (Krauss and Nyblade, 2005). Party leaders have also strengthened their hold over public aspects of their parties. For example, since the end of the 1980s, Belgian party leaders have increasingly become the sole spokespersons of their parties (Fiers and Krouwel, 2005). The minimal evidence available to date provides empirical support for the existence of this kind of centralized behavioural personalization.
Decentralized behavioural personalization of politicians refers to a process in which individuals – e.g. candidates or members of parliament – who are not party or executive leaders, increasingly engage in individual activities and step back from involvement as team players in their group (political party, cabinet). Empirical evidence of decentralized behavioural personalization is hard to find. In fact, we have none, save for the case of Israel, which will be discussed below and which lends support to the claim that such a personalization trend occurs.
Behavioural personalization of voters
Behavioural personalization of voters implies that people vote more on the basis of their evaluations of individuals and identification with them, and less on the basis of evaluations of parties and identification with them.
Centralized behavioural personalization of voters implies that people vote more on the basis of their evaluations of leaders and less on the basis of evaluations of parties and identification with them. In the US, Wattenberg (1991) found that political leaders have become electorally more important in their own right by personifying the policy platforms of their respective parties (see also Huang and Price, 2001; Keeter 1987; McAllister, 1996, 2007). A similar trend was found in Australia and New Zealand, both parliamentary democracies (Bean, 1993).
On the other hand, findings in other studies demonstrate that voters do not cast their vote based on the identity of the party leader to a greater extent than in the past. For example, Curtice and Holmberg’s (2005) study of party leader effects in elections in six Western European countries (Denmark, Germany, Great Britain, The Netherlands, Norway and Sweden) found that the independent effect of leader evaluation was very limited and showed no increase over time. Similarly, King’s (2003) study, which examined six cases (US, Britain, France, Germany, Canada and Russia) over a period of 40 years, concluded that although there were a few cases where specific leaders made a crucial difference, leaders were generally not decisive for election outcomes (for further evidence, see also Hayes, 2009; Holmberg and Oscarsson, 2004; Kaase, 1994). Therefore, while the empirical evidence regarding this type of personalization is extensive, we find it to be mixed in respect of the level of support for its occurrence.
Decentralized behavioural personalization of voters implies that voters vote more on the basis of their evaluations of individual candidates (not specifically the party leaders), and less on the basis of their evaluations of parties and their identification with them. Karvonen (2009) found this trend in some – but not all − countries with semi-open list electoral systems, as voters increasingly tend to exploit the opportunity of exercising their personal preferences. He also argued – on the basis of his examination of nine countries (Belgium, Denmark, Sweden, Finland, Ireland, The Netherlands, Germany, Britain and Norway) – that ‘in countries where the possibility of choosing between individual candidates has existed for a long time, the relative importance of individual candidates seems to have increased’ (p. 28). In The Netherlands, for example, the party is still the main influence on voters, where only 8 percent are ready to follow their preferred candidate if the candidate decides to run under the auspices of another party. Yet there is evidence of a trend of decentralized personalization given that an increasing percentage of voters cast their single personal vote for candidates who are not party leaders (van Holsteyn and Andeweg, 2010). Thus, there is a moderate amount of empirical evidence for this type of personalization with mixed levels of support for its existence.
Centralized and decentralized personalization: Israel as a case study
We believe that the lack of empirical evidence supporting or refuting the occurrence of various types of personalization, and especially of various types of decentralized personalization, results from a conceptual muddle. Rahat and Sheafer (2007) took a first step towards clarification by suggesting the typology of personalization used above (institutional, media and behavioural), while this article takes another step adding the distinction between centralized and decentralized personalization. We use the Israeli case to demonstrate the usefulness of this new typology. We employ various indices to detect trends of personalization that scholars might utilize in their studies of other countries.
Israel provides an ideal setting in which to examine personalization trends. It is a parliamentary democracy that experienced the heyday of mass political parties and these parties have been the dominant actors in the Israeli polity (Galnoor, 1982; Horowitz and Lissak, 1989). Thus Israel supplies a party-dominated starting point for measuring the rise of individual-centred politics. Additionally, as mentioned above, a general trend of personalization in Israel has been identified empirically in previous studies (Rahat and Sheafer, 2007; Shenhav and Sheafer, 2008).
Institutional personalization in Israel
The two main features of the Israeli government: the parliamentary system and the closed list, single nationwide district electoral system, are party-centred. Yet studies have demonstrated that institutional personalization does occur in Israel and is sometimes seen as a way to bypass these fundamental, party-centred features (Rahat, 2008).
Some incremental enhancement of the prime minister’s powers has occurred over the years, such as granting authority to dismiss ministers in 1981. Direct elections of the prime minister − a reform adopted in 1992 and implemented from 1996 until its abolition in 2001 – was a clear acceleration of this trend of centralized institutional personalization. The repeal of direct elections did not necessary signify a reverse trend. Following the repeal, a number of features were adopted that enhanced the power of the prime minister. These included the prime minister’s ability to call for early elections (with the president’s consent) and the determination that a government could be ousted in a vote of no confidence only in cases in which there was majority support for an alternative leader to attempt to form a government. These changes significantly expand the prime minister’s power, while direct elections for prime minister weakened him in practice (Hazan, 2005). All in all, over the past three decades, we have witnessed a continuous process of centralized personalization.
In Israel, we also witnessed an example of the combined impact of centralized and decentralized personalization when, due to the adoption of party primaries in 1992 by one of the largest parties, and in 1996 by another, the apparently powerful (directly elected) prime minister lost control over his party. Decentralized institutional personalization at the intra-party level led to personalized behaviour of ministers and representatives (Hazan and Rahat, 2000).
At the intra-party level, reforms in leadership and candidate selection methods have also led to centralized and decentralized personalization starting in the 1970s and culminating in the 1990s. The formalization of leadership selection and the adoption of more inclusive selectorates − first party delegates and then party members − in the large parties (Kenig, 2009b) imply that power has become more invested in party leaders. That is, leaders are more autonomous in policy-making (Katz and Mair, 1995) and are more secure in their positions (Kenig, 2009a). At the same time, however, candidate selection methods have also been reformed (Barnea and Rahat, 2007). This decentralized personalization is expressed in the development of an explicit personal intra-party competition and in the perception of a personal mandate among party representatives. Moreover, success in intra-party competition is now key in becoming a minister (Kenig and Barnea, 2009).
Concerning candidate selection, an additional different trend of personalization has emerged over the past five years. In several parties, a self-selected leader him/herself selects the party’s candidates for election to the Knesset, or at least is highly influential in their selection. Thus, if up to 1996, we could for the most part speak of a trend toward the opening up of candidate selection methods in terms of decentralized personalization, today we can also point at a trend toward taking power away from the team (be it the party central committee or the party congress) and bestowing it upon party leaders (Rahat, 2011). In other words, if until recently developments in candidate selection were in the direction of decentralized personalization, a reverse trend towards centralized personalization is also evident in some parties over the last five years.
A clear symbol of centralized personalization in general elections for the legislature is the addition of the party leader’s name to the ballot (Figure 1). This is a development in campaign communication signifying centralized personalization within the legal constraints of the party-centred, closed list system. This trend began at the end of the 1970s and the beginning of the 1980s in small parties which were formed by single leaders, but it broke into the mainstream when the Labour Party added the name of its popular leader, Yitzhak Rabin, to the ballot in 1992. Before that, only the name of the party and a series of letters that signified the party appeared on the ballot paper.
The number of party lists that included the name of their leader in their official name in the elections, 1949–2009.
Institutional personalization is also prominent in local government. Israel adopted a reform in 1975 which replaced the old parliament-like local council with direct mayoral elections. In the past, citizens elected the members of the council through a proportional representation list system in which the municipality was a single district, and the head of the municipality was in turn selected by the council, requiring the support of the majority to sustain her position. Following the reform, the local council is still elected according to the old system but the heads of local governments are elected directly by the people and cannot be ousted and replaced by the council. This reform centralized personal power in the hands of the mayor. Over time – due to split ticket voting and the holding of local elections separately from the national elections – the new system also led to the decline of most parties at the local level, to the point that most members of the council act independently rather than as members of a team. The reform at the local level directly implied centralized personalization but also led over time to (behavioural) decentralized personalization.
Media personalization in Israel
Most data appearing in this section are based on content analysis of media coverage of 16 election campaigns for the Knesset (the Israeli parliament) that took place in Israel between 1949 (the first elections) and 2003 (the last elections that were analysed). 1
Personalization in the unpaid media
An increase in media centralized personalization is evident when looking at developments over time of the Israeli media’s coverage of elections. Figure 2 demonstrates that over the years there has been a decline in the proportion of news items that focus on the party (exclusively or combined with the party leader) and an increase in the proportion of news items that focus on party leaders.
Centralized personalization in media election coverage.
When it comes to decentralized personalization, the expectation is that politicians who are not party leaders will also become more prominent. Looking at available evidence, this trend is not as apparent as that relating to centralized personalization. Our analysis shows that − save for a peak in 1969 − there has been no clear overall trend of either decline or increase in the coverage of party candidates who were not party leaders (data not shown). 2 Overall, when looking at all campaigns from 1949 through 2003, between 50 and 70 percent of all news items covering elections referred to candidates who were not the party leaders. These high percentages might be interpreted as indicators of personalized politics, but a positive trend, which is at the core of the definition of personalization, is not evident. Figure 3 demonstrates, however, that within the news items that did mention junior party candidates who were not party leaders (N = 1,135), 3 a trend towards decentralized personalization is evident. Specifically, leaving aside items in which party leaders appeared and analysing only those mentioning junior party candidates and the party itself, it is apparent that the centrality of the junior candidates in comparison with the party has increased over time. When items in which party leaders are also mentioned are included in the analysis, this trend disappears, probably because the centralized personalization process overshadows the decentralized personalization process.
Decentralized personalization in election coverage: Media focus on junior party candidates.
Centralized personalization is also apparent in news coverage of another collective entity, the Israeli government. Figure 4 presents the ratio between the number of times that the government was referred to using the name of the prime minister (e.g. the Sharon government) and the number of times in which it was called the ‘Israeli government’. We see that during the first three decades, the ratio stood at less than 1:10; between 1977 and 1996 it increased to a ratio of 1:5, and between 1997 and 2007 it jumped to a ratio of almost 1:2. In other words, the government is identified more and more with its leader than with the nation it is supposed to serve.
Centralized personalization in news coverage of the government.
Personalization in the paid media
Figure 5 points to a clear trend towards centralized personalization in party advertisements. Over the years, newspaper advertisements have focused more on leaders and less on the party, and there is a similar increase in the percentage mentioning the leader by name. Yet, as the figure shows, parties still appear more prominently than their leaders in this kind of campaign communication. This comes as no surprise because in each election – except for the 1996 and 1999 elections which included direct election of the prime minister – voters in Israel have been asked to vote for a party, not for a person.
Centralized personalization in parties’ campaign communication.
When focusing on junior candidates in party advertisements, we see mixed results concerning decentralized personalization. Figure 6 shows that on the one hand there is a decline in the percentage of party candidates mentioned in party newspaper advertisements. At the same time, there is an increase in the number of advertisements including photographs of party candidates.
Decentralized personalization in campaign communication.
Personalization in the behaviour of Israeli politicians
A possible proxy for behavioural personalization is the ratio between the number of governmental positions (ministers and deputy ministers from all coalition parties) and the total number of Members of the Knesset (MKs) in the coalition. The idea is simple: stronger parties will establish and maintain a relatively small government because they can ensure the loyalty of their parliamentary representatives in the coalition. In a personalized environment, loyalty would be bought on a personal basis and the government would thus be expected to be larger. Figure 7 clearly shows a constant decrease in the ratio between the number of ministers and deputy ministers and the size of the coalition (r = –0.70). It is evident that the price of coalition support has increased over the years, from one position for every six coalition members to one position for less than two coalition members. 4 Over the years, more positions have had to be distributed in order to form and maintain coalitions. This can be understood to result from a decline in party cohesion and in their ability to discipline representatives. In other words, the personal rewards required to build and maintain coalitions have become inflated, since loyalty is bought on a personal basis in return for ministerial posts and additional posts in parliament. An extreme example of this is the number of posts given to the Labour Party in the 32nd government, formed in April 2009. Labour held 13 Knesset seats. Five of their Knesset members received ministerial posts, two became deputy ministers and one became the faction head, while the remaining five acted and voted quite independently. That is, the support of seven of the eight members of Labour who supported the 32nd government was obtained in exchange for executive posts. We are thus witnessing decentralized personalization in this respect. Note that this type of decentralized personalization is at the same time about the government and party leaders. Like the prime ministers, who are also leaders of their parties, coalition party leaders require more positions for their own people in order to ensure that the coalition is maintained.
Ministers and deputy ministers per coalition member.
Another sign of decentralized personalization is the increased number of private members’ bills submission and adoption. Studies of Israeli parliamentarism have shown that the number of private members’ bills submitted to the Knesset rose from less than ten in each of the first two Knessets (1949–1955); to dozens in the next two (1955–1961); to hundreds in the next seven terms (1961–1988); and to thousands since then (Blander and Klein, 2002). The percentage of legislation in each Knesset term originating in private members’ bills has also grown. In the first ten Knessets (1949–1984), excluding two minor and temporary declines, the share of legislation that originated in private members’ bills grew from 2 percent in the 1st Knesset (1949–1951) to 21 percent in the 9th (1977–1981). The 1984–2003 period saw a sharp upsurge in legislation originating in private members’ bills from 16.2 percent in the 11th Knesset (1984–1988) to 59.3 percent in the 15th (1999–2003). 5
The increased use of roll-call voting, in which the name of each member of the Knesset is called and the vote is recorded in the Knesset protocols, can also be seen as a sign of decentralized personalization. The number of roll calls was less than ten for each five-year period up to 1972. It increased consistently in each of the next five-year periods, starting with 19 roll-call votes in the 1973–1978 period and ending with 71 roll-call votes between 1996 and 2002 (Rahat, 2007).
An additional indication of decentralized personalization in the behaviour of MKs is the sharp rise in the number of petitions they submit to the High Court of Justice (Dotan and Hofnung, 2005). At the same time, ‘a reverse process has taken place in relation to petitions submitted by parties’ (ibid., p. 89). This increase is explained as an expression of the wishes of MKs to gain public exposure that they need, especially after the adoption of party primaries by several parties (institutional decentralized personalization).
Another indication of behavioural decentralized personalization in Israel is the dramatic increase in the number of interjections directed by MKs to the prime ministers’ speeches on peace initiatives (Shenhav, 2008). Heckling and interjections by MKs have been identified as a discursive mechanism to shift the focus of audiences, and mainly the media, from the content of prime ministers’ messages to the scene taking place in the plenum.
The interface between legislators and the head of the executive branch also supplies examples of centralized personalization. From a discursive point of view, self-references in this context, for example the use of the first person singular, is a possible proxy to measure personalization. 6 An empirical examination of this type of discursive personalization is presented in Figure 8 summarizing use of the first person singular in prime ministerial speeches at the introduction of new governments from 1950 to 2009. Figure 8 shows a clear indication of personalization over the 60 years of Israeli statehood. The use of the first person singular climbs from being hardly used to no less than 1.5 percent during the 2000s.
First person singular per total words in speeches, 1949–2009.
The collective memory of Israeli politics seems to support this trend. Rabin’s victory speech following the Labour victory in the 1992 elections became a milestone for personalization of Israeli political discourse: ‘I will navigate the coalition negotiation … I will decide who will be the ministers’ (cited in Azulay-Katz, 1992: 2). Prime Minister Netanyahu also used the navigation metaphor, declaring as he presented his government to the Knesset: ‘I will be the one to navigate Israel’s economic strategy’ (Netanyahu, 2009). A similar tone was also set by Prime Minister Sharon, who repeatedly referred to ‘my’ disengagement plan (Barnea and Sheafer, 2004; Sharon, 2004a, b and c). 7 This discursive personalization undertaken by prime ministers is yet another example of centralized personalization.
Behavioural personalization of voters in Israel
There is evidence that supports the claim for the occurrence of centralized behavioural personalization of voters. There has been an increase over time in the percentage of respondents in surveys who mention the party leader as their determining factor in voting for a party, with a simultaneous decline in the percentage of respondents who claim that they make their electoral decision based on identification with a particular party. Supporting evidence can be seen in Figure 9, where it is demonstrated that over a period of six elections Israeli voters generally voted more on the basis of their evaluation of leaders and less on the basis of their party identification. 8
Centralized behavioural personalization of voters.
No comparable empirical evidence is available for decentralized behavioural personalization of voters, probably because this concept has hardly been discussed in previous studies of elections.
Discussion
The main contribution of this article to the growing scholarly attention to political personalization in recent years is the conceptual and theoretical development of the distinction between centralized and decentralized personalization and the demonstration of its utility for empirical analysis.
This article suggests that political personalization implies changes in power distribution in two possible directions: on the one hand, it can foster a centralized dynamic in which more power flows up to the leadership, while, on the other, it can imply decentralization, which diffuses power from groups to the individuals of which various political groupings are composed. This means that the balance of power is significantly changing, flowing up (to party leaders, prime ministers) and down (to individual politicians within each party or to cabinet members). In both cases the losers in these processes are the teams, parties and cabinets, and the gainers are individual politicians. Such processes do not balance or offset each other; rather they imply a highly significant change, especially for parliamentary democracies whose starting point is more party-centred and whose basic institutional structure is built on collectives, parties and cabinets.
Personalization implies a decline in the role of parties – a decline that is likely to be pronounced in some or all of the functions performed by political parties: People identify with personalities rather than with parties; individual politicians, rather than parties, become the representatives of specific policies; interest aggregation occurs more on an ad hoc basis rather than within parties; individuals rather than parties communicate with the public; policy emerges from interaction between individuals in government rather than as a product of debate and deliberation within the party; and, to a certain extent, candidates and leaders select parties rather than the other way round.
From a liberal standpoint, this process might be viewed positively. Power is now spread more evenly among leaders, elites and individual politicians, and this allows them to check and balance each other. But it also might be seen as a negative trend toward a plebiscitary model, a Napoleonic mode of democracy: a dominant leader speaks directly to the people or to an atomistic crowd of individual politicians and cannot be balanced by the weakening political groupings, institutions and organizations. Which view one adopts of the influence of personalization on democracy is a function of one’s view of the starting point. If one thinks that political groupings had an appropriate amount of power in the past, personalization will be an issue of grave concern. If one thinks that political groupings were too powerful, one might see this process positively in terms of distribution of power.
In either case, we believe that there is much to be concerned about, particularly with regard to the rapidity of the personalization processes that, at least in the Israeli case, are taking place simultaneously at various fronts. Changing the balance between groups and individuals may affect the mechanisms that limit and constrain the power of leadership. While individuals at the top are empowered, political groupings are weakened and are replaced with strengthened individual politicians. Yet these individuals act on their own so they cannot do much to counterbalance their leaders, especially because they lack the backing of other members of the declining groupings.
Since we are dealing with democracies, there are many other features that will restrain the leader, so the emergence of a dictatorship is not an immediate danger. Moreover, more personally powerful leaders will be more accountable because their source of power is still the citizens’ vote and their personal prominence only makes them more susceptible to personal failure and defeat, as well as triumph. As Webb and Poguntke (2005: 346) put it: ‘Presidentialized party leaders are less likely to survive electoral defeats than their precursors.’ The fear is, however, that leaders will fail to lead, one after another, because they lack the necessary mechanisms to build and sustain support and legitimacy from their political groupings. In the long run, this gap between the seemingly empowered leaders with a decreased ability to deliver – since they have to aggregate interests on an individual rather than on a group basis – could decrease public trust in democracy. This threat is especially problematic in cases of party-centred systems. In such cases, accelerated processes of personalization might foster incongruence between regime structure and processes of legitimization. A coexistence of legal tradition and institutional structures of group-centred politics and processes of personalization produce internal tensions that might endanger a leadership’s ability to gain legitimacy. Instead, constant conflicts between norms, institutions and political behaviour might threaten the system’s ability to sustain the in-depth processes of political legitimization so crucial for democracies.
The adaptive nature of democracies raises expectations that they will either adjust to centralized and decentralized personalization, or will strive to construct new or renewed collective identities and groupings. Whatever the case may be, the centralized and decentralized effects of political personalization will be of major concern.
Footnotes
Acknowledgements
The article was presented at the Joint Israeli–Canadian research workshop, ‘The Institutionalization of Political Parties in Canada and Israel: Cohesion, Discipline and Personalization’, the Halbert Center of Canadian Studies at the Hebrew University of Jerusalem, June 2011.