![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
Abstract
Are labour market outsiders also political outsiders? The labour market dualization literature offers inconsistent evidence as to the effects of nonstandard employment on political demobilization. This article shows that thus far unaccounted for exposure to poverty among workers in nonstandard employment varies considerably across labour markets, and that this variation carries implications for political inclusion. Analyses of five waves of the European Social Survey indicate that exposure to working poverty absorbs much of the predictive power of nonstandard employment, suggesting that working poverty, rather than labour market per se, leads to political demobilization. The findings help explain the hitherto inconsistent results in studies of labour market dualization and lend support to the argument that future research should account for cross-national variation in working poverty.
Acknowledgements
The author would like to thank Eva Anduiza, Macarena Ares, Enrique Hernández, Marga León, Lara Maestripieri and Tim Hellwig for helpful comments on an earlier draft. This research has was funded with a Marie Sklodowska Curie Individual Fellowship (EU project 742991, LABOREP).
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Workers in nonstandard dependent employment are those on part-time and temporary contracts. Workers in standard employment are those on full-time stable contracts. For reference, I examine apart upscale professionals, the self-employed, the unemployed and those inactive on the labour market. My classification follows Emmenegger (Citation2009).
2 A more recent strain of research estimates ‘risk exposure’ to nonstandard employment based on demographic and occupational characteristics (Rovny and Rovny Citation2017; Schwander and Häuserman Citation2013). This approach, too, fails to address cross-national heterogeneity in economic conditions between workers in nonstandard employment.
3 In some cases I had small subgroup samples. A missing value was assigned for the risk-of-poverty estimate where N < 200. The threshold was set at 200 because annual estimates below the threshold deviated considerably from previous years. The results were replicated with a more stringent cut-off, at 300, and the conclusions presented here hold.
4 Data from EU-SILC covers the twelve months prior to each survey.
5 Levels of working poverty defined with the seven-month minimum are highly correlated with a stricter definition of only full-year workers and with a looser definition of all persons who have worked at least one month (Lohmann Citation2018).
6 The measure accounts for cross-nation differences in cost of living as expressed in purchase parity standards. The 60 percent threshold is widely used since the implementation of the Laeken indicators, the Eurostat’s official indicators of social inclusion and poverty.
7 Horemans et al. (Citation2016) show that this is due to a drop in the income of workers in nonstandard employment.
8 Working poverty is therefore fundamentally different from unemployment or a low-wage rate which reflect the labour market position of a single worker (Maitre et al. Citation2011; Marx and Nolan Citation2014).
9 The self-employed are a diverse group that cannot be subsumed as either in standard or nonstandard employment (Jansen Citation2019). Due to data limitations, the false self-employed could not be distinguished from the traditionally self-employed.
10 The wide confidence interval for upscales in Slovenia is due to the small sample size (N = 26). With the sample threshold of 200 observations, upscales in Slovenia are coded as missing.
11 For the unemployed, the samples range from 382 individuals in Sweden to 3,904 individuals in Spain.
12 This is akin to other risk-based measures in the literature, such as occupational unemployment rates or outsiderness (Iversen and Soskice Citation2001; Rehm Citation2011; Schwander and Häusermann Citation2013).
13 The variance at the labour market level is reduced from 0.195 in Model 1 to 0.122 in Model 2, with poverty risk as a predictor, a reduction of 37 percent.
14 Häusermann et al. (Citation2018) show that macroeconomic conditions interact with education in their impact on turnout. I apply a fixed-effects specification and leave it to future research to disentangle how contextual factors condition the impact of poverty risk on turnout.
15 Due to endogeneity issues I have discarded a model that includes all interactions and their constituent terms. Namely, estimates of poverty risk are informed by – and thus endogenous to – labour market status. Endogeneity issues become apparent in the models presented in the Online Appendix (Model 10 in Table A.2). The size of the coefficients increases substantially, and so do the standard errors, both of which are tell-tell signs of multicollinearity.
Additional information
Notes on contributors
Dani M. Marinova
Dani M. Marinova is Serra Hunter Professor in Political Science at Universitat Autònoma de Barcelona. Her research focuses on political behaviour and political economy. Her articles have appeared in Perspectives on Politics, Political Behavior, and Political Science Research and Methods. Her book Coping with Complexity: How Voters Adapt to Unstable Parties (ECPR Press) won the 2017 GESIS Klingemann Prize for Best CSES Scholarship. [[email protected]]