3  Conclusion

The aim of this study was to explore the social foundations of the transformation of Swiss capitalism since the 1980s. The rationale was to show that the neoliberal transformation of the Swiss hybrid model had to be founded on a strategy of institutional change supported by a sufficiently broad social base.

The crisis of the 90s did not mark a break-up of the Swiss dominant social bloc, but rather a re-composition and a change in the power balance within the bloc itself. While the crisis did not destabilized the left bloc that much, the right bloc, that I named “liberal conservative” throughout this study, underwent important process of change.

In fact, the social bases of the political parties representing the liberal conservative bloc, the PRD-PLR, PDC and UDC, changed substantially since the 80s. Unlike Durrer de la Sota, Gethin, and Martinez-Toledano (2021), who showed that the social base of right-wing government parties stayed rather stable since the 70s, the results presented in the previous section show that, while the UDC’s social basis was rather merchant in the 80s, it transformed into a popular basis during the 90s. Furthermore, the social basis of the PRD (then PLR) used to be very bourgeois until the 2010s, period after which this social base showed trends towards an evolution into a merchant right. Interestingly, my results are more supportive of the brahmin left vs merchant right divide hypothesis than the WPID’s own results for Switzerland. However, if one looks at political leaning instead of party support, the left appears to be bourgeois in the sense that it is supported by the upper-middle income groups and by university graduates, while self-identification to the right evolved into a merchant right pattern.

Then, an analysis closer to the methods and conceptual framework of the neorealist approach was carried out through a latent class analysis for the late 1990s in order to assess more closely the re-composition of the left and liberal conservative blocs during the crisis. The estimated LCA model divides Swiss population of the late 90s in twelve socio-political groups. The twelve clusters model shows that the bourgeois basis that, mediated by the PRD, could support a neoliberal and pro-EU strategy was rather narrow (around 5% only). In fact, a strategy aggregating social groups which were opponents of EU integration, conservative vis-à-vis immigration and cultural issues would gather a broader base. This strategy was endorsed by the UDC, which became the first Swiss party in terms of vote shares and gained a second seat in the federal council after the 2003 national elections.

As stated in the introduction, the present work is a first attempt to employ neorealist framework and methodology to study institutional change in Switzerland. The present work suffers from several limitations that could be improve in further work. First, the time span ranges from 1981 to 2016. Although the time frame can’t really go back further than 1981, it could be extended to most recent years. This could be done, for instance, by analyzing the VOTO studies (the successor of the VOX studies since 2017) which collected post-voting survey data from 2017 to 2020. This analysis of the VOTO studies could be complemented, as done in this work, by a latent class analysis using more recent data, for instance the SELECTS 2019 panel study, which offers even better questions than the SELECTS post-election data for 1999 used in this work. Overall, more analyses of more recent data would be a good complement to the present study to assess the stability of the new social bloc configuration which emerged in the 90s.

Finally, the present work tends to overlook the crucial role of ideology, political strategies and institutional change. On the one hand, the different political strategies aimed at promoting a new Swiss socio-economic model, both on the left and the right, deserve to be more carefully explored. On the other hand, ideology, both at the Swiss and international levels, could explain the transformations of these strategies, but also the transformation of the social basis of the different blocs. Moreover, the process of institutional change that the Swiss socio-economic model underwent since the 90s deserve more careful analysis. All these points could be the subject of further work.

4 Appendix

Figure 4.1: Party identification

Figure 4.2: Proportions of Swiss citizen leaning for the left, the right or neither left or right: 1988-2016

Figure 4.3: Proportions of Swiss citizen supporting state intervention, equal rights & chances for foreigners, and redistribution: 1993-2016

Table 4.1: Binary logit regressions: party support (log-odds). Full table available here

	PS	Greens	No party
incomefhigh	-1.054***	-1.407*	-0.251*
	(0.181)	(0.568)	(0.104)
incomefmiddle_high	-0.309**	-0.555+	-0.289***
	(0.113)	(0.313)	(0.077)
incomefmiddle_low	-0.012	-0.361	-0.082
	(0.102)	(0.278)	(0.070)
educationfcomp_school	-0.083	-1.244***	0.202**
	(0.103)	(0.367)	(0.068)
educationfhigher_voctraining	0.507***	0.249	-0.412***
	(0.137)	(0.335)	(0.102)
educationfmaturity	-0.345	1.296**	-0.629***
	(0.228)	(0.398)	(0.163)
educationfUniversity	-0.156	1.302**	-0.651***
	(0.261)	(0.406)	(0.188)
annee_cat(1986,1991]	-0.919***	0.327	0.287**
	(0.137)	(0.309)	(0.092)
annee_cat(1991,1996]	-0.601***	0.211	0.119
	(0.120)	(0.295)	(0.083)
annee_cat(1996,2001]	-0.940***	-1.191**	0.248**
	(0.124)	(0.372)	(0.082)
annee_cat(2001,2006]	-0.667***	-0.295	-0.118
	(0.118)	(0.305)	(0.082)
annee_cat(2006,2011]	-0.751***	0.148	-0.237*
	(0.143)	(0.319)	(0.101)
annee_cat(2011,2016]	-0.645***	-0.169	-0.355***
	(0.143)	(0.350)	(0.103)
Num.Obs.	68501	68501	75456
AIC	60477.9	18126.4	96941.1
BIC	61208.6	18857.2	97679.6
Log.Lik.	-30158.935	-8983.191	-48390.555
F			64.224
RMSE	0.37	0.17	0.47
+ p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001

Table 4.2: Binary logit regressions: party support (log-odds). Full table available here

	PRD-PLR	PDC/PCS	UDC
incomefhigh	1.262***	0.267	-0.008
	(0.165)	(0.179)	(0.209)
incomefmiddle_high	0.969***	0.068	-0.033
	(0.142)	(0.143)	(0.165)
incomefmiddle_low	0.382**	0.005	0.023
	(0.139)	(0.133)	(0.153)
educationfcomp_school	0.129	0.034	0.012
	(0.130)	(0.124)	(0.135)
educationfhigher_voctraining	0.301	-0.344	-0.672**
	(0.189)	(0.211)	(0.212)
educationfmaturity	0.442+	-0.420	-1.473**
	(0.241)	(0.312)	(0.461)
educationfUniversity	-0.014	-0.306	-1.505*
	(0.287)	(0.333)	(0.607)
annee_cat(1986,1991]	-0.698***	-0.619***	-0.670**
	(0.187)	(0.171)	(0.228)
annee_cat(1991,1996]	-0.417**	-0.737***	-0.064
	(0.161)	(0.159)	(0.182)
annee_cat(1996,2001]	-0.668***	-0.970***	0.171
	(0.165)	(0.160)	(0.172)
annee_cat(2001,2006]	-0.575***	-1.039***	0.965***
	(0.164)	(0.164)	(0.161)
annee_cat(2006,2011]	-0.316+	-0.597**	0.947***
	(0.189)	(0.188)	(0.185)
annee_cat(2011,2016]	-0.281	-1.132***	1.267***
	(0.185)	(0.200)	(0.183)
Num.Obs.	68501	68501	68501
AIC	46836.1	34929.2	41848.2
BIC	47566.8	35660.0	42579.0
Log.Lik.	-23338.037	-17384.620	-20844.110
RMSE	0.31	0.27	0.30
+ p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001

Table 4.3: Binary logit regressions: political leaning (log-odds). Full table available here

	Left	Right	No leaning
income[3k,5k]	-0.238+	0.256+	0.006
	(0.129)	(0.136)	(0.117)
income[5k,7k]	-0.003	0.151	-0.094
	(0.133)	(0.143)	(0.123)
income[7k,9k]	-0.014	0.368*	-0.259+
	(0.150)	(0.159)	(0.143)
incomemore_9k	-0.068	0.540***	-0.380*
	(0.157)	(0.162)	(0.151)
educationfcomp_school	-0.258+	0.238	-0.003
	(0.150)	(0.145)	(0.125)
educationfhigher_voctraining	0.370*	0.081	-0.368*
	(0.160)	(0.169)	(0.148)
educationfmaturity	0.975***	0.090	-1.049***
	(0.185)	(0.211)	(0.197)
educationfUniversity	0.908***	0.070	-1.019***
	(0.142)	(0.157)	(0.151)
annee_cat[1996,2000)	-0.245+	0.079	0.127
	(0.133)	(0.141)	(0.120)
annee_cat[2000,2004)	-0.493***	0.102	0.298*
	(0.135)	(0.142)	(0.120)
annee_cat[2004,2008)	-0.146	0.204	-0.044
	(0.134)	(0.144)	(0.123)
annee_cat[2008,2012)	-0.468**	0.275+	0.158
	(0.155)	(0.161)	(0.139)
annee_cat[2012,2016]	-0.335*	0.490**	-0.099
	(0.147)	(0.151)	(0.133)
educationfmaturity × annee_cat[2012,2016]	-0.462**	-0.033	0.584***
	(0.165)	(0.183)	(0.177)
educationfUniversity × annee_cat[2012,2016]	0.033	-0.378**	0.352*
	(0.127)	(0.135)	(0.138)
Num.Obs.	45278	45278	45278
AIC	52978.1	52450.1	58592.7
BIC	53693.2	53165.1	59307.8
Log.Lik.	-26407.061	-26143.026	-29214.367
RMSE	0.45	0.44	0.48
+ p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001

Table 4.4: Binary logit regressions: issue position (log-odds). Full table available here

	Redistribution	Equal R&C CH-foreign	State intervention
income[3k,5k]	0.002	-0.022	-0.112
	(0.103)	(0.105)	(0.116)
income[5k,7k]	-0.100	0.099	-0.329**
	(0.109)	(0.111)	(0.125)
income[7k,9k]	-0.362**	0.158	-0.220
	(0.126)	(0.128)	(0.145)
incomemore_9k	-0.650***	0.217	-0.400*
	(0.134)	(0.136)	(0.157)
educationfcomp_school	0.162	-0.307**	0.246*
	(0.110)	(0.111)	(0.121)
educationfhigher_voctraining	-0.180	0.330*	-0.128
	(0.137)	(0.136)	(0.160)
educationfmaturity	-0.011	0.952***	-0.016
	(0.169)	(0.173)	(0.188)
educationfUniversity	0.100	1.276***	0.347*
	(0.129)	(0.137)	(0.139)
annee_cat[1996,2000)	-0.023	-0.154	0.015
	(0.106)	(0.108)	(0.120)
annee_cat[2000,2004)	-0.105	-0.003	0.397***
	(0.104)	(0.106)	(0.116)
annee_cat[2004,2008)	0.051	0.167	0.508***
	(0.108)	(0.110)	(0.119)
annee_cat[2008,2012)	0.033	0.079	0.696***
	(0.124)	(0.125)	(0.132)
annee_cat[2012,2016]	0.123	-0.076	0.599***
	(0.118)	(0.119)	(0.127)
educationfmaturity × annee_cat[2012,2016]	-0.528***	-0.607***	0.102
	(0.152)	(0.155)	(0.171)
educationfUniversity × annee_cat[2012,2016]	-0.210+	-0.276*	0.221+
	(0.117)	(0.124)	(0.129)
Num.Obs.	52418	52434	52537
AIC	71130.4	69248.4	64698.3
BIC	71857.5	69975.5	65425.6
Log.Lik.	-35483.212	-34542.213	-32267.140
RMSE	0.49	0.48	0.46
+ p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001

Figure 4.4: Average marginal contrasts results from multilevel probit model: party support

Figure 4.5: Average marginal contrasts results from multilevel probit regressions: political leaning

Figure 4.6: Average marginal contrasts results from multilevel probit regressions: issue position

Table 4.5: Model selection: 12 clusters model

Cluster	L²	L2 bootstrap p	-2LL Diff	-2LL Diff bootstrap p	Entropy R²
1-Cluster	14058.41	0.000	NA	NA	1.0000
2-Cluster	12111.01	0.000	1947.3954	0.000	0.6441
3-Cluster	11173.20	0.006	937.8165	0.000	0.6299
4-Cluster	10932.57	0.060	240.6254	0.000	0.6043
5-Cluster	10708.48	0.192	224.0898	0.000	0.5925
6-Cluster	10584.08	0.234	124.4011	0.000	0.5788
7-Cluster	10467.74	0.324	116.3432	0.000	0.5933
8-Cluster	10389.37	0.356	78.3647	0.000	0.6009
9-Cluster	10318.10	0.332	71.2689	0.000	0.6084
10-Cluster	10259.10	0.290	59.0079	0.004	0.6058
11-Cluster	10208.58	0.264	50.5185	0.014	0.6035
12-Cluster	10160.79	0.264	47.7854	0.034	0.6107
13-Cluster	10114.68	0.268	46.1100	0.066	0.6293

Figure 4.7: Variables chosen as indicators in the Latent class model for 1999: descriptive statistics

Table 4.6: Binary logistic regression: support for EU integration. Log(odds). Data source: SELECTS 1999

	For EU integration
income	0.013***
	(0.003)
education	0.007*
	(0.003)
as_character(sex)male	-0.032+
	(0.018)
as_character(language)German	-0.139***
	(0.021)
as_character(language)Italian	-0.240***
	(0.027)
symp_udc	-0.052***
	(0.003)
symp_ps	0.039***
	(0.003)
symp_prd	0.016***
	(0.004)
Num.Obs.	2365
R2	0.301
AIC	2530.4
BIC	2588.1
Log.Lik.	-1255.195
RMSE	0.41
+ p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001

Durrer de la Sota, Carmen, Amory Gethin, and Clara Martinez-Toledano. 2021. “Party System Transformation and the Structure of Political Cleavages in Austria, Belgium, the Netherlands and Switzerland, 1967-2019.” World Inequality Lab – Working Paper, no. 2021. https://wid.world/wp-content/uploads/2021/03/WorldInequalityLab_WP2021_08_PoliticalCleavages_AustriaBelgiumNLSwitzerland.pdf.