Abstract:

Many scholars argue that education or occupation, not income, are key to explain electoral realignment. Yet income is generally conceptualized as short-term, current income. In this paper, I introduce “permanent income”, a long-term income measure over people’s entire lifespan, to re-examine the relationship between income and electoral realignment. I argue that permanent income closely overlaps with education and that the association between permanent income and voting is driven by important changes in composition and size of education-income groups. Using national election studies from Denmark, the United Kingdom, and the United States from the 1950s to the present, I find that electoral realignment is driven by higher-educated voters with high permanent income increasingly associated with mainstream left parties, and by lower-educated voters with low permanent income increasingly associated with mainstream right parties. The findings point to both material and non-material channels behind the effects of income and education on electoral behaviour.

Working paper (September 2022) available on request.

Abstract:

Abstract:

In this paper, we explore the transformative political and policy implications of artificial intelligence (AI), a key driver of the most recent wave of rapid technological innovation. Our focus is twofold: we aim to enhance our theoretical understanding of AI and its unique characteristics, and we provide empirical evidence on AI’s real-world impact in firms from both managerial and employee perspectives.

Our research questions include: how do managers and employees perceive AI and its impact on the workplace? What social (unemployment) policies and regulation do managers and employees prefer to deal with AI’s (potential) impact on unemployment and inequality? To what extent does information on potential job losses due to AI affects managers’ and employees’ preferences for unemployment policy and regulation?

The central hypothesis we explore is that managers and employees have systematically different estimations of the impact of AI, even within the same firm or sector. We furthermore hypothesize that the more the economic effects of AI are portrayed as negative, the more both employers and employees will support compensatory unemployment policy.

We contribute, first, by dissecting AI from the broader concept of digital transformation, such as automation and robotization, to better define its unique features and risk profile. We highlight how AI challenges traditional notions of the direction, speed, and ‘bias’ of technological change, necessitating theoretical innovation.

Second, we present empirical findings from a novel firm-level survey of Danish managers, coupled with a register-based survey of employees. This combined approach allows us to study the actual impact of AI on individual companies from complementary perspectives. Specifically, we investigate managerial and employee estimations of AI’s impact on employees, how AI affects employers’ and employees’ social policy preferences, and (experimentally) to what extent information on the amount of potential job losses due to AI leads to a change in preference for unemployment policy and regulation. 

Abstract:

Descriptive misrepresentation is widespread, but surprisingly persistent. Why are affluent and highly educated voters over-represented in politics, even among populist radical right parties? In this paper, we explore how descriptive and symbolic forms of representation interact and appeal to voters. We argue that political elites can engage in symbolic representation through class signaling, which can compensate for descriptive misrepresentation. Using original data from a conjoint survey experiment in Switzerland in 2023, we find that voters are generally biased in favor of party leaders from working-class backgrounds and class-neutral cultural behavior. More importantly, we demonstrate that both types of symbolic class signaling can turn voters from under-represented groups (non-university-educated, women) in favor of party leaders from over-represented groups (university-educated, men). Hence, symbolic representation can compensate for descriptive misrepresentation. This contributes to our understanding of the puzzle why descriptive misrepresentation persists.

Abstract:

The slowdown of economic growth and the stagnation of incomes for substantial parts of the population in recent decades are well-known. But what are the implications of these changes for the politics of taxation? The consensus in the literature is that income change either has no effect or that large income decline raises support for welfare policies. Focusing on the revenue side of the welfare state rather than the spending side, we present the opposite argument: We predict that economic decline makes individuals less tolerant of paying taxes, because tax increases would imply a reduction of their consumption level. We test this argument using longitudinal data from both repeated cross-sections and panel surveys in the United States, Canada and Japan. Our main finding is that tolerance of paying taxes is lower when individuals perceive that their economic situation deteriorates. Thus, perceived economic decline can create political obstacles against higher taxation.

Download working paper here (June 2022).

Abstract

What is the effect of income stagnation on support for the welfare state? Combining novel data on the evolution of income to existing micro- and macro-level datasets, we argue that stagnation leads to greater support for spending cuts and tax cuts. We develop a simple model linking income stagnation to support for cuts via three distinct mechanisms. Stagnation reduces altruistic motives for welfare state spending, it heightens the relative perceived costs of insurance, and it leads individuals to support tax cuts to compensate for their stagnating incomes. Our micro-level empirical analyses show that individuals facing stagnant or lower incomes support spending cuts and tax cuts to a greater extent. This effect is especially strong among high-income individuals and/or people facing low unemployment risk. At the macro level, these dynamics lead to greater retrenchment in countries with lower income growth. Taken together, our findings link the literature on income stagnation to comparative political economy studies of changing welfare states. They help us make sense of why governments implement (often) economically inefficient spending cuts during economic crises, despite rising risks which should lead to larger welfare states. In contrast to previous literature claiming that the implementation of austerity is the result of democracy being subverted, we show that there are rationally based reasons why some pivotal electoral groups may support retrenchment under conditions of economic stagnation.

Download working paper here (July 2022).

Abstract

While classical models in political economy suggest that citizens facing adverse economic situations prefer to increase redistribution and welfare state effort, recent studies have shown that economic decline is associated with a reduction of willingness to pay taxes. Therefore, economic decline makes citizens more likely to want “something for nothing”: they don’t accept to pay for the additional public spending they demand. We test our arguments in a hard case, Norway, where support for taxes and spending are generally coherent and ideologically determined. Finding that Norwegians facing economic hardship want additional spending, but lower taxes would go against the consensus in the literature on public opinion on welfare states in Nordic countries. Our empirical analysis uses three waves of a panel survey conducted in Norway in 2014, 2015 and 2017 which allows us to leverage variation between and within individual respondents. We study the impact of both perceived and objective economic risks on two measures of preferences for welfare state benefits, two measures of willingness to pay taxes and a joint measure of support for taxes and spending. We find that low income and low education voters as well as those with high perceived economic risks are more likely to prefer something for nothing: they are in favour of more welfare spending, but tend to oppose tax increases.

Abstract:

This paper examines the relationship between liberalization, which involves removing market barriers and restrictions on free markets, and income inequality in 18 democratic capitalist countries from 1976 to 2013. We argue that the distributive impact of liberalization depends on the specific characteristics of the policies implemented, such as the policy field and intensity of the reform. The study uses a new dataset on liberalization reforms and finds that liberalization reforms are associated with higher levels of income inequality in the long run, primarily through widening the distribution of market incomes and reducing redistribution by the state. We show that liberalization affects bottom-end inequality more strongly than top-end inequality and that liberalization of taxes and transfer policies are more strongly associated with rising inequality than other policy areas. Finally, we provide evidence for a compensation effect, where de-liberalization policy reforms can partially moderate the inegalitarian impact of liberalization reforms.

Abstract:

This study shows that voting behavior in European countries can be explained by peoples’ stance on economic inequality, in particular their fairness evaluation of the income distribution. The argument is that unfairness perceptions trigger either a desire for more equality or socio-economic resentment. A desire for equality arises because people strive for a fair distribution of economic resources, and it is associated with a higher (lower) probability of voting for left (right) parties due to their equality-related policy stances. Resentment arises especially when people think that they themselves are treated unfairly, and it comes with feeling of being “left behind” unduly as well as a wish to boost their social status. Resentment is associated with voting for radical parties, especially on the right, because they offer a discourse and policies that aim to boost the standing of the “unduly disadvantaged”. Our empirical analysis based on European Social Survey (ESS) data supports these theoretical ex-pectations. Unfairness perceptions are strongly associated with a higher probability of voting for the left and a lower probability of voting for the mainstream right. Further analyses suggest that these effects are mediated by redistribution preferences (issue-based voting for the left or against the right) and political distrust, dissatisfaction, political efficacy, and immigration attitudes (resentment-based voting for the radical right).

Abstract: