Abstract
Does becoming poorer always cause people to shift their attitudes towards higher demand for redistribution? Through a systematic review of the literature on this question, we reveal five important themes in existing research: a person’s current income, their future expectations, their expectations about redistribution benefits, their income in early life and their attitudes towards beneficiaries. Identifying these themes helps explain why responses to economic hardship are variable and heterogeneous, and can very usefully guide future research.
Does becoming poorer always cause people to shift their attitudes towards higher demand for redistribution? This systematic review focuses on the literature linking people’s experiences of economic hardship to their support for redistribution, diverting government support towards the poorer in society, usually funded through progressive taxation of the rich. Based on a review of over 100 articles and books, mainly in the political economy literature, we identify five key themes that moderate the individual-level relationship between economic hardship and redistribution attitudes, helping open useful new directions for empirical work. We focus on a person’s current income, how much they expect to earn in future, whether they think they gain from higher welfare spending, the social context in which they grew up and their attitudes towards people and groups seen to benefit most from government support. 1
Our review highlights several implications of this nuanced and wide-ranging literature. The first is that responses to economic hardship are heterogeneous, and vary systematically according to the themes we identify. The second is the identification of research gaps, especially around public perceptions of government spending and its beneficiaries, and the effects of time horizons and expectations of financial recovery. The third is for policymakers and political practice: public support for greater redistribution is not an inevitable consequence of losses to people’s economic wellbeing, but depends on the factors we outline in this review.
Definitions and Difficulties
There are many ways to define ‘redistribution attitudes’, with each leading to different interpretations. Is support for redistribution best captured by advocating greater spending on the poor specifically rather than greater spending overall, or is it more about the principle of taxing wealthy people more to promote equality? We know, for example, that it matters who spending is diverted from, and diverted towards. Our first insight, then, is that ‘support for redistribution’ is not one concept and there is no consensus on any one measure in this literature. This variation is important when interpreting different findings, and helps us identify research gaps in assessing the state of knowledge. Researchers investigating support for redistribution need to carefully consider how measures map onto concepts, and the resulting interpretation of their meaning.
A related challenge is defining the types of economic difficulties that increase support for redistribution. Any negative income shock (or drastic increase to inflation that reduces a person’s economic wellbeing) leads to a downward trajectory that can sometimes shift social attitudes (Helgason and Rehm, 2023). Our reading of this literature reveals the usefulness of distinguishing between three groups: the stationary rich, the stationary poor and those who become poor. For the former, we expect the significance of an income shock to be relatively modest, given that it will likely make less difference to the life they are able to lead. For the ‘stationary poor’, we would expect support for redistribution to start at a relatively higher level (on average), if they expect to benefit from redistribution, and so a further income shock may not substantially change that level of support. However, those who enter into hardship, and those on a further downward trajectory, should, we propose, be most likely to change their attitudes according to the main observations in this literature. This group becomes aware of new circumstances, and so demands, and potentially relies on, state support.
These observations highlight the importance of subjective economic evaluations. This separation of perception and reality has received pushback, because subjective economic evaluations might simply reflect prior attitudes (Evans and Pickup, 2010; Rueda and Stegmueller, 2019; Weisstanner and Armingeon, 2022). However, omitting perceptions misses very useful insight when those perceptions are the valid sources of more pro-redistribution views. Economic perceptions can be relativised by one’s social network, for example, with people feeling richer or poorer than they are, objectively, in response to the wealth of those they associate with (Cansunar, 2021), a phenomenon psychologists term ‘social sampling’ (Dawtry et al., 2015; Galesic et al., 2012). Many middle-income individuals are ‘over-leveraged’ in costly mortgages or other credit markets (Kumhof et al., 2015), introducing a perception of hardship, despite relative wealth. And common cognitive biases, like how people experience financial losses asymmetrically to gains (loss aversion) or place greatest emphasis on recent events (recency bias), permeate the political effects of economic change (Gilovich et al., 2002; Tversky and Kahneman, 1973).
Having summarised the broad definitional and scope issues on this topic, we proceed to summarise five themes identified through our systematic review. We first consider material explanations (Themes 1, 2 and 3) oriented around one’s income and how redistributive policy measures might cause it to change. Each theme is illustrated with a question an individual might ask themselves in their response to a material worsening of their financial circumstances, sufficient – potentially – to change their support for redistribution.
Theme 1: What Is My Income Now?
The foundational theoretical framework for most material explanations of redistribution preferences is that of Meltzer and Richard’s (1981) study, a minimalist theory of taxation and public spending that extends work by Romer (1975). In this stripped-down account, people are driven by their ‘pocketbook’ as it is right now, and they will support public spending if they believe its benefits outweigh the cost of higher taxes. Considering a flat tax rate – one that is constant across all incomes – voters with mean (average) income see their ‘pocketbook’ perfectly balanced between the benefits of redistribution and the cost of taxation. In most countries income is right-skewed, such that the median (middle) person, and hence the majority, falls below the mean and would rather increase the tax rate to receive greater support. As a result, aggregate support for redistribution will strengthen as the median income falls further below the mean. If the majority of the population will go on to vote for redistributive parties in elections, we have a mechanism by which public spending – which they refer to as the size of government – increases. This mechanism is dependent on an economically self-interested ‘rational’ individual. This is broadly, if indirectly, consistent with evidence showing that public support for increased spending rises in periods of welfare reduction, and falls in periods of increased spending – the ‘public thermostat’ model of public opinion for government spending (Wlezien, 1995).
Setting aside these aggregated predictions, which find surprisingly little empirical support (Ansell and Samuels, 2014: chap. 1), we focus on the model’s insights about individual-level behaviour. Here there is stronger evidence, with income and support for redistribution being negatively correlated (Rueda and Stegmueller, 2019). However, one puzzling result is that significant numbers of middle-class and wealthy people support higher levels of government spending, while significant proportions of society’s poorest want less. Against the expectations of Meltzer and Richard (1981), Hoy and Mager (2021) show that distributional information reduces support for redistributive spending across 10 countries, an effect driven by those who previously over-estimated their relative wealth. Some have also posited the ‘last-place aversion’ hypothesis, with relatively poor people not wanting to be the poorest (Kuziemko et al., 2014). Therefore, while there is evidence of a broad link between income and support for redistribution, this relationship is not perfectly correlated nor deterministic, and its mechanisms leave discrepancies to be explored.
Theme 2: How Do I Think I’ll Do in Future?
One way scholars have resolved this tension is to loosen the focus on present-day income and instead to think about the longer term. For some, redistribution is insurance against the risk of future financial losses; while two workers might earn an identical wage, the worker whose job is less secure might support higher social spending to protect against the elevated risk of unemployment (Iversen and Soskice, 2001; Thewissen and Rueda, 2019). This logic considers an individual’s ability to act with a longer time horizon in mind. It can also help resolve some empirical puzzles. Rehm (2009) outlines how overall support for redistribution and public spending is maximised when income and risk are negatively correlated; the poor support redistribution because of their present income, the rich through fear of one day losing their well-paid job.
Building on this, Rueda and Stegmueller (2019) explicitly model individuals’ expected income trajectory, using age and education as proxies for what an individual will earn over their life. They find that these future expectations explain a significant portion of variation in redistribution preferences. Langsæther et al. (2022) find similar when studying class mobility in Britain, another proxy for potential future earnings.
These results are important, since expectations condition the effect of income losses and may explain why people of different backgrounds, but with equivalent present income, respond differently to shocks. Some economic hardship produces great uncertainty over a significant period of time, leading to a substantial re-evaluation of expected lifetime earnings that drive a change in preferences. But a comparable shock that is more transient should have only trivial effects on long-run expectations. It is within this framework that we can also situate studies that look at wealth as well as income. Ansell (2014), for instance, shows that increasing house prices – a measure of household wealth – have negative effects on support for redistribution.
Theme 3: How Much Will I Actually Benefit from Government Spending?
The above frameworks are underwritten by a ‘nonregressivity assumption’ (Iversen and Soskice, 2006); richer voters lose out from redistribution, while poorer voters benefit. But despite its theoretical ramifications, this assumption is less well founded, and might be more variable than first thought (Beramendi and Rehm, 2016; Holland, 2018).
To see why, we can consider redistributive public spending to have de jure and de facto beneficiaries, those who the policy makes better off on paper versus those who actually stand to gain. On the former, existing work on welfare states has highlighted implicit middle-class biases in how public spending is targeted. Heavy government spending on or subsidisation of childcare, education, transport infrastructure and even tax-free allowances benefit educated ‘sociocultural professionals’ – such as teachers, doctors, civil servants and so on – more than they do members of the conventional working class (Gingrich and Häusermann, 2015). Perhaps because of these patterns, many poorer voters might not perceive themselves as beneficiaries of government spending, current or proposed, and will hence have less material reason to support redistribution.
Relatedly, a push towards welfare conditionality, and the introduction of strict sanctions, has made it harder for the poor to access services to which they are entitled. O’Grady (2022) shows how waiting times for appointments, the necessity of evidencing job applications and a general presumption of wrongdoing have significantly increased the effort required for people to access services. As scholars have shown in a variety of contexts, increased hostility and bureaucracy make people less supportive of welfare measures, less likely to think they benefit from them and less likely to feel well represented or trusting of government (Campbell, 2012). Taken together, they might perceive minimal personal gain from welfare spending, and so have less incentive to support redistribution than is often assumed.
Broader Considerations
We now consider explanations not rooted in this cost–benefit logic of social behaviour. These do not disregard material frameworks, but complement them.
Theme 4: What Was My Income in Early Life?
The fourth theme considers the effects of income in one’s early years. Socialisation-based accounts of preference formation contend that people’s early formative years shape how they think about politics for the rest of their lives. The role of parents, family and young adulthood is crucial, with lifestyle, attitudes and beliefs filtering through in a manner that profoundly, and somewhat permanently, shapes one’s worldview (Sears and Funk, 1990).
Neundorf and Soroka (2018) highlight how unemployment and precarity when one first enters the labour market (usually in a person’s early twenties) have lasting pro-redistributive effects. Shorrocks and Grasso (2020) make a similar argument about gender differences in redistribution support, while O’Grady (2019) uses panel data from Switzerland to show the predictive power of parental income. In a lab experiment, Almås et al. (2017) show that family income and socioeconomic status are drivers of fairness beliefs among teens.
For O’Grady (2019), early experiences are not quite as permanent as certain proponents of socialisation might suggest (see Helgason and Rehm, 2023). Rather, socialisation creates a frame of reference against which future events are benchmarked, leading to rational decision-making that is ‘bounded’, where only highly noticeable income shocks have the power to substantively shift one’s preferences. This explains not only why income and preferences exhibit such striking cross-sectional differences, arising through socialisation and gradual long-run shifts in income, but also why evidence of individuals quickly changing their preferences, in response to particular shocks, is more elusive (Margalit, 2019).
This reasoning helps explain, in part, why only large income shocks may have effects on preferences for redistribution and why a person’s baseline attitudes determine their subsequent responsiveness. Margalit (2013), for example, finds that job losses have larger, more positive effects on the redistribution preferences of Republicans than of Democrats in the US. Republicans have a lower baseline support, for partisan and ideological reasons, so – counterintuitively – have greater room for manoeuvre.
Theme 5: How Much Do I, and Others, Deserve to Benefit?
Finally, we focus on a key belief system that might pass from one generation to the next; a disposition to believe society is meritocratic, and the extent to which anyone deserves support from the state (Alesina and Angeletos, 2005; Alesina and Glaeser, 2004). Such debates are motivated in the morality of taking from the wealthy to help the poor, an action that can be criticised if one believes that income is proportional to effort (Trump, 2020). There is evidence that Norwegians – socialised into a highly egalitarian society – tend to see inequality as more ‘unfair’ in behavioural experiments than American counterparts (Almås et al., 2020). The wealthiest 5% of Americans are more likely to say their wealth stems from personal choices rather than luck (Suhay et al., 2021).
Different ideological dispositions might explain variation in the acceptance (Mijs, 2021; Trump, 2020) and awareness (Waldfogel et al., 2021) of social inequalities, and why exposure to them can lower support for redistribution (Sands, 2017).
One area of work highlights differences between wealthier and poorer people in responding to different facets of redistributive policies. Cavaillé and Trump (2015) show that attitudes towards which groups should receive support are not explained by income, and it is better accounted for by the literature on social affinity. Rueda (2018) proposes that the wealthy support redistribution when inequality is high and ethnic diversity is low, on the grounds that increased poverty and homogeneity increase the perceived deservingness of support. Moreover, people may even resist spending that they could also benefit from, if ‘out-groups’ also gain from those shared resources (Breznau and Eger, 2016; Mitchell, 2019; Shayo, 2009). Hansen (2023) argues that public perceptions of the rich are also critical, with support for higher taxation lessened when they are seen to contribute to society and help others, rather than being greedy and self-interested (Hansen, 2023; McCall, 2013).
There is compelling evidence that support for redistribution rises when the poor are seen to exert effort, but is curtailed when they are seen to freeride off others (Aarøe and Petersen, 2014). Large parts of these beliefs might stem from deep-rooted ideological worldviews, with evidence that politicians and the media play a role too (O’Grady, 2022).
Arguments about the deservingness of government support have been tied to literature on social discrimination. Empirically, aggregate support for redistribution declines when individuals believe that minority groups are its primary beneficiaries and are disproportionately taking scarce resources (Alesina and Glaeser, 2004; Baldwin and Huber, 2010; Gilens, 2000). These studies point to the importance of group perceptions and deservingness as being intrinsically tied to support for redistribution. This reinforces the importance of perceptions of ‘who benefits’ from redistribution – both lower income groups who do not expect to gain (discussed under Theme 3) and higher income groups’ impression of likely beneficiaries.
Conclusion
This systematic review points towards five main themes which, collectively, can explain the link between economic hardship and support for redistribution. These are illustrated through five questions an individual may consider, according to the culmination of research to date: What is my income now? How do I think I’ll do in the future? How much will I personally gain from government spending? What was my income in early life? And how much do I, or anyone else, deserve to benefit? These themes provide a useful starting point for deeper empirical and conceptual work going forwards.
The first question (‘what is my income now?’) points to a need for greater focus on how people view their current income and their current economic circumstances. The question, ‘how do I think I’ll do in future?’ is, as highlighted earlier, a function of both long-term returns (Rueda and Stegmueller, 2019) and perceived risk (Rehm, 2009). Researchers might investigate different types of economic and income shocks, some which may induce greater feelings of risk and threat, under some conditions rather than others, such as whether a person has other ‘insurance’ to fall back on (the wealth of family members, a perceived inheritance, etc.). With respect to risk, the perception of employment risk, as well as pension and returns (and so on), might be a useful occupation-specific area of investigation. Our third question (‘how much will I actually benefit from government spending?’) gets to the heart of the core assumption that poorer individuals will demand higher redistribution. This link – between harder financial circumstances and demands for greater state support – has within it an intermediate step that has not, to our knowledge, been investigated, namely the assumption that poorer people necessarily benefit of government spending and investment.
Moving beyond cost–benefit style frameworks, our question ‘what was my income in early life?’ relates to the importance of socialisation and benchmarking on later experiences and expectations. Longitudinal studies would be valuable here, alongside deeper questions in cross-sectional or subsequent panel studies of the economic experiences of respondents in their formative years. Finally, we turned to attitudes around deservingness and likely beneficiaries (‘how much do I, or anyone else, deserve to benefit, and who will benefit?’) as key conditions for greater support from the state. Studies should explore who people think are the beneficiaries of state spending. Further consideration of attitudes about the ‘deserving’ and ‘undeserving rich’ in different contexts and economic climates is the important area for further work.
As we have shown in this review, public support for greater redistribution is not an inevitable consequence of losses to people’s economic wellbeing. There is considerable heterogeneity in response to income shocks, for individuals and nationally, when there are inflationary shocks or recessions. Different sub-groups should be expected to react differently, according to perceptions about their income and long-term expectations, beliefs about who benefits from government spending, childhood financial and political circumstances and attitudes towards ‘in’ and ‘out’-groups. As we noted at the start of this review, studying these conditionalities remains highly sensitive to questions of measurement, both of redistribution preferences itself and of people’s starting points, changes and trajectories of economic experience.
There are implications, too, for politics and policymakers. Understanding redistribution preferences, and their responsiveness to economic hardship, underpins sustainable policy solutions which seek to reallocate resources with the legitimacy of public support. With this in mind, we return to our previous discussion of eligibility for government support. While minority groups are typically poorer, and hence may often receive more redistributive support from the state, it is also true that many aspects of government spending benefit all groups in society, or at least have the potential to. Understanding people’s beliefs about the beneficiaries of redistribution and government spending seems, particularly in light of the potential for some universality of benefits to gain greater and broader support, on the basis of our review, to be a promising time and important area for future work.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research from Nuffield College, University of Oxford.
