Chiocchetti ECPR Class voting in Italy
Download as PPTX, PDF
0 likes118 views
This document analyzes class voting patterns in Italian elections from 1987 to 2018. The author finds that class is no longer a strong predictor of voting behavior, with explained variation being medium-low and the direction of the class-party relationship often incorrect. Parties have shifted their class bases, with the left-wing PD becoming more upper-class oriented and the populist Five Star Movement behaving like a lower-class party. These findings have implications for theories of linear class-party associations and suggest the need for alternative conceptualizations of class coalitions in voting.
Chiocchetti ECPR Class voting in Italy
- 1. Class voting in Italy: a new realignment? Dr Paolo Chiocchetti Postdoctoral Researcher Humboldt-Universit辰t zu Berlin paolo.chiocchetti@hu-berlin.de
- 2. Outline 1. Introduction 2. Results: class voting in Italy, 1987-2018 3. Implications
- 3. The impact of class on political and voting behaviour is one of the classic areas of political research Lipset & Rokkan 1967; Bartolini 2000; Alford 1962; Manza & al. 1995; Evans 1999; Van der Waal & Achterberg 2007 The theory of class voting narrowly defined rests on two hypotheses: 1. societies are divided into hierarchical social classes which show differential patterns of voting behaviour e.g. relations of production (Marx, Rokkan), control over social resources (Weber), occupational categories (Erikson & Goldthorpe 1992, Rose & Harrison 2010) 2. there is a coherent and stable class-party association: upper categories tend to support right-wing parties, lower categories tend to support left-wing parties Some problems (theoretical, methodological and empirical) Empirical research on Italy: literature divided on the continued relevance of class voting after 1992 Bellucci 2001; Ballarino et al 2009; Maraffi 2018 1. Introduction
- 4. Puzzle: who is right? I will revisit the empirical evidence to find out Data: Italian National Election Studies (ITANES) datasets for the 1987, 1996, 2006 and 2018 elections intermediate elections to be filled in later, previous elections not entirely Improved class scale based on relations of production (primary) and occupational status (secondary) 1. self-employed (and employers) 2. cadres (and managers) 3. teachers 4. skilled employees 5. unskilled employees 6. manual workers 7. unemployed based on individual positions (inactive population excluded)
- 5. Testing 1.series of linear regressions for each party and year coefficient of determination (R2): fitness of the model regression coefficient (m): direction and strength of class voting 2. visual inspection of the class voting curve for each party and year Expectations (ideal case) 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0% Self-employed Cadre Teacher Skilledemployee Unskilledemployee Manualworker Unemployed Left-wing party Right-wing party Left-wing party R2: 100% (perfect) m: +11% (pos. & str.) Right-wing party R2: 100% (perfect) m: -11% (neg. & str.)
- 6. The empirical analysis does not confirm the importance of class voting in Italy (except for 1987) 1. Medium-low explained variation (R2): model not very explanatory 2. Small strength of effect (m): effect of class scale on party vote small 3. Direction of effect frequently wrong! 2. Results Election R2 (non directional) m (non directional) correct direction 1987 24.5 1.5 100.0 1996 29.4 0.8 41.9 2006 33.7 0.7 67.8 2018 36.0 1.0 40.1 WEIGHTED AVERAGE 28.4 1.0 62.5 Table 1. Results by election (in %)
- 7. Party Type Avg. share R2 m (normalize d) correct direction PCI (1) left 23.6 40.1 +16.7 100.0 PRC (3) left 4.7 25.7 +1.7 66.7 PD (3) left 18.3 22.7 -1.5 33.3 PSI (1) left 12.0 3.3 +3.3 100.0 PRI-PSDI-PLI- DINI (2) mixed 5.6 33.1 -11.4 (100.0) OTHER (4) mixed 9.8 25.9 -6.7 (0.0) POP (1) unclear 5.3 7.4 -6.3 (0.0) DC (1) right 25.9 18.6 -4.1 100.0 UDC (2) right 5.2 57.3 -10.9 100.0 FI (3) right 14.5 4.7 -0.4 33.3 LN (3) right 8.4 2.1 -0.4 33.3 FdI (3) right 7.9 14.7 -4.9 100.0 M5S (1) - 23.1 80.5 +12.0 (100.0) Abstention (4) - 20.1 55.2 +13.1 (100.0) Table 2. Results by party (in %)
- 8. EXAMPLE 1: PD (left) progressive shift from lower-class (PCI 1987) to upper-class party (PD 2018) 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% 40.0% Self-employed Cadre Teacher Skilledemployee Unskilledemployee Manualworker Unemployed PCI 1987 PDS 1996 Ulivo 2006 PD 2018
- 9. EXAMPLE 2: FI (right) does not behave like an upper-class party (U-shaped behaviour) 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% Self-employed Cadre Teacher Skilledemployee Unskilledemployee Manualworker Unemployed FI 1987 FI 1996 FI 2006 FI 2018
- 10. EXAMPLE 3: FIVE STARS MOVEMENT (populist) behaves like a lower-class party 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% Self-employed Cadre Teacher Skilledemployee Unskilledemployee Manualworker Unemployed M5S 1987 M5S 1996 M5S 2006 M5S 2018
- 11. EXAMPLE 4: NON-VOTERS behave like a lower-class party 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% 40.0% Self-employed Cadre Teacher Skilledemployee Unskilledemployee Manualworker Unemployed Abstention 1987 Abstention 1996 Abstention 2006 Abstention 2018
- 12. More analysis needed on explaining the causes of these shifts 1) declining relative salience of class cleavage probably true, but cannot explain much in itself 2) occupational categories not reflecting class relations possible 3) changing class nature of Italian parties (ideology, policies, organization) likely, see M5S in 2018 4) class-based coalitions are not linear also likely Serious implications for the theory of class voting 1) traditional class-party association (lower-class & left, upper-class & right) cannot be assumed to be a valid hypothesis 2) focus on contingent class coalitions rather than on a dichotomy or scale 3) devise alternative measurements not based on 3. Implications