Introduction
The magnitude and speed of the economic collapse that followed the outbreak of COVID-19 were unlike any experienced in a century. Governments and central banks responded with fiscal, monetary, and financial market interventions that were unprecedented in size, scope, and swiftness, providing vital lifelines to households and businesses.
These measures speak to important questions about social insurance and redistribution. Existing studies of COVID-19-related economic policy interventions have focused mainly on public attitudes toward social protection policies and have generated conflicting findings. Some find that personal experiences of the pandemic shifted attitudes toward policy interventions, albeit in different directions: toward social investment policies benefiting working-age citizens and away from old-age pensions in Germany, Sweden, and Spain (Enggist et al., Reference Enggist, Häusermann and Pinggera2022); toward social protection for more affluent Germans impacted by the pandemic (Tonelli et al., Reference Tonelli, Harris and Deeg2024); and away from government intervention in Australia due to sustained restrictions on personal freedom (Biddle et al., Reference Biddle, Edwards, Gray and McAllister2024). Other studies find that public attitudes remained broadly stable during the pandemic (Ares et al., Reference Ares, Bürgisser and Häusermann2021; de Vries et al., Reference de Vries, Geiger, Scullion, Summers, Edmiston, Ingold, Robertshaw and Young2023; Ebbinghaus et al., Reference Ebbinghaus, Lehner and Naumann2022; Kim, Reference Kim2024; Weisstanner, Reference Weisstanner2022).
We contribute to this literature by taking a wider perspective on social insurance and redistribution during crises. The policy response of governments and central banks in virtually all advanced economies over 2020–2021 included two main elements: social insurance of employment/income risk via wage subsidies and financial-sector interventions that stabilized household, corporate, and financial sector balance sheets. Among advanced economies, fiscal measures and central bank balance sheet expansions were both more than double the size of those undertaken during the first year of the global financial crisis (GFC) (Hooley et al., Reference Hooley, Khan, Lattie, Mak, Salazar, Sayegh and Stella2023; IMF, 2020).
We assess how fairness considerations and policy design shaped attitudes toward both kinds of intervention using a conjoint experiment embedded in an online survey undertaken in Australia and the United Kingdom (UK) in early 2021. The conjoint survey shows that respondents were broadly supportive of the interventions, but that support for wage subsidies was significantly higher than for financial and credit market interventions. Drawing on scholarship on deservingness and fairness (Cappelen et al., Reference Cappelen, Konow, Sørensen and Tungodden2013; Gualtieri et al., Reference Gualtieri, Nicolini and Sabatini2019; C. Jensen and Petersen, Reference Jensen and Petersen2017), we argue that shared narratives about the origins of the COVID crisis and intuitive fairness beliefs about ‘reciprocity’ and ‘proportionality’ induced strongest support for wage subsidies (Atari et al., Reference Atari, Haidt, Graham, Koleva, Stevens and Dehghani2023; Cavaillé, Reference Cavaillé2023). The COVID pandemic could readily be understood either as having distributed losses randomly or as imposing the largest losses on relatively vulnerable citizens. Either way, employment and income losses could be seen as mostly outside individuals’ control and of a different kind than that experienced during a ‘normal’ business cycle, heightening perceptions that those experiencing them were deserving of social insurance. Policy design features were also of critical importance. Wage subsidies had strict, widely publicized eligibility requirements mostly tied directly to individuals’ pre-pandemic employment, minimizing the risk that they might be perceived as encouraging free riding. These subsidies were notably less generous toward recent immigrants and temporary residents, further mitigating concerns about free-riding. However, respondents who are more right-wing or who perceive that the level of inequality is lower or not increasing remained less supportive of wage subsidies (Alesina and Giuliano, Reference Alesina and Giuliano2009; Stantcheva, Reference Stantcheva2021).
Unlike other studies (Ferwerda et al., Reference Ferwerda, Magni, Hooghe and Marks2024; Haverland et al., Reference Haverland, van der Veer and Onderco2022; Malmendier, Reference Malmendier2021), we also find that attitudes toward pandemic wage subsidies were little affected by pandemic experiences or by material and social divides, including income, wealth, and ethnicity. This contrasts with classic findings on redistribution preferences (e.g. Alesina et al., Reference Alesina, Baqir and Easterly1999a).
Regarding financial and credit market interventions, we find mostly supportive attitudes but at significantly lower levels than for wage subsidies. Notably, interventions supporting large firms and banks were more opposed. We argue that this is best explained by the greater difficulty many people have in viewing interventions benefiting large banks and firms as consistent with key fairness norms. These companies are more readily perceived as being opportunistic, benefiting from political connections, and as less genuinely in need than small firms and households relying on employment income (Cavaillé, Reference Cavaillé2023; Culpepper et al., Reference Culpepper, Shandler, Jung and Lee2025; Gourinchas et al., Reference Gourinchas, Kalemli-Ö zcan, Penciakova and Sander2021). Financial interventions are also more opaque and less well understood by the public, which constrains their appeal on fairness grounds. Thus, support for pandemic economic interventions varied and policy design matters.
In the next section, we outline our theoretical motivations and hypotheses. The third section describes our method and data. The fourth presents our results and the final section concludes.
Theory and hypotheses
Economic policy interventions in the COVID-19 pandemic included extraordinary wage subsidies to those unable to work, payment holidays for personal debts and rental contracts, assistance to small and large businesses, and central bank interventions, including interest rate deductions, liquidity provision, and asset market stabilization measures (IMF, 2021). Taken together, these measures supported household and business incomes and balance sheets.
Our primary argument is that fairness beliefs and policy design shape variations in public support for these economic policy interventions. We argue that these factors are likely to induce greater public support for wage subsidies than for financial-sector interventions, including payment contract relief and central bank policies. We also outline an alternative theory that people’s policy preferences were determined primarily by varying personal experiences of the pandemic. Finally, we discuss several individual-level attributes that could shape policy preferences and for which we control in our analysis.
Fairness beliefs and policy design
An important strand of research suggests that people’s attitudes toward public policy interventions are stable and associated with beliefs about merit and inequality (e.g., Alesina and Glaeser, Reference Alesina and Glaeser2004; Alesina et al., Reference Alesina, Stantcheva and Teso2018; Duval et al., Reference Duval, Ji, Papageorgiou, Shibata and Spilimbergo2024; Feldman and Steenbergen, Reference Feldman and Steenbergen2001; Fund, Reference Fund2024; Stantcheva, Reference Stantcheva2024). In this perspective, preferences about economic policy interventions are linked to people’s judgments about their alignment with understandings of fairness and deservingness (Cappelen et al., Reference Cappelen, Konow, Sørensen and Tungodden2013; Cassar et al., Reference Cassar, Healy and von Kessler2017; Gualtieri et al., Reference Gualtieri, Nicolini and Sabatini2019; Heap et al., Reference Heap, Koop, Matakos, Unan and Weber2020; C. Jensen and Petersen, Reference Jensen and Petersen2017; Petersen, Reference Petersen2012).
The viral origin of the crisis likely triggered innate moral intuitions about care, equal treatment, and purity (Atari et al., Reference Atari, Haidt, Graham, Koleva, Stevens and Dehghani2023). These and other intuitions plausibly also shaped how people thought about government policy responses, heightening the perception that recipients of wage subsidies were deserving of assistance. Cavaillé (Reference Cavaillé2023) argues that fairness reasoning rests on two intuitive foundations: a reciprocity norm whereby all group members contribute collectively to social insurance and draw benefits when in real need, and a proportionality norm whereby market outcomes should match individual effort. Especially among the relatively wealthy, tolerance for free-riding tends to shrink during recessions.
Deviations from fairness as reciprocity occur when people believe that what Cavaillé refers to as ‘redistribution to’ benefits free riders. Public health interventions by governments generated employment income losses of a different kind than ‘normal’ business cycle unemployment. The risks posed by viral exposure and lockdowns requiring self-isolation were likely viewed by many as outside the control of those experiencing them (Cappelen et al., Reference Cappelen, Konow, Sørensen and Tungodden2013; Gualtieri et al., Reference Gualtieri, Nicolini and Sabatini2019; C. Jensen and Petersen, Reference Jensen and Petersen2017). By imposing widely distributed, quasi-random losses, these effects could have produced a general sense of a commonly shared exogenous shock, reducing concerns that income protection would benefit the undeserving (Cappelen et al., Reference Cappelen, Konow, Sørensen and Tungodden2013; Cassar et al., Reference Cassar, Healy and von Kessler2017; Gualtieri et al., Reference Gualtieri, Nicolini and Sabatini2019; Heap et al., Reference Heap, Koop, Matakos, Unan and Weber2020; Petersen, Reference Petersen2012).
In such circumstances, more people are likely to perceive wage subsidies as fair and necessary to prevent undeserved economic hardship. Moreover, their more widespread availability to affected individuals outside the long-term unemployed – despite their unusual generosity – lessened their redistributive implications. The exceptional character of the pandemic shock could also have reduced concerns about the fiscal sustainability implications of the measures, at least in the initial phases of the crisis.
Although people can have high-level beliefs about fairness and how it relates to misfortune and inequality of outcomes, their judgments are likely to depend on policy specifics. Policy design shapes expectations about who, when, why, and how interventions benefit or disadvantage specific groups (Jordan, Reference Jordan2013; Larsen, Reference Larsen2008; Svallfors, Reference Svallfors2012. Policy-designs during the pandemic reflected considerable innovation. Wage subsidies were very generous by historical standards in Australia and the UK, approaching levels provided by European social democracies in the case of the UK. This could have heightened public concerns – shared by incumbent governments of the center-right in both Australia and the UK – that they could benefit less deserving people and be unaffordable.
However, in both countries, wage subsidies were universally available yet conditional on strict, widely publicized eligibility requirements, including verified COVID-related employment disruption. They were delivered through businesses rather than by government welfare agencies – in contrast, for example, to the United States, which gave cash transfers to individuals. Although both countries also introduced wage replacement schemes for self-employed workers, these had strict eligibility and reporting requirements, including submission of tax returns and minimum trading periods. This signaled to voters that the potential for free-riding among the self-employed would be minimized.
In the UK, the Coronavirus Job Retention Scheme (CJRS, commonly referred to as ‘furlough’) provided social insurance of employment/income risk via grants to firms to pay 80% of workers’ salaries and employment costs, up to a total of £2,500 per month. The CJRS was initially to run for only three months, but was first extended until October 2020, then to April 2021, and finally until the end of September 2021. The level of the wage subsidy fell to 70% in September 2020 and 60% in October, before increasing back to 80% from November 2020 as the UK tightened restrictions during a second national lockdown. The level again fell to 70% in July 2021 and 60% in August and September. In total, 11.6 million jobs were furloughed in the 18 months of the scheme (Pope and Hourston, Reference Pope and Hourston2022). The linking of subsidy generosity to the intensity of lockdown restrictions also probably reassured the public that the level of subsidy was connected with involuntary unemployment and that it would be temporary.
The Australian government implemented a similar scheme in March 2020 after a series of one-off payments to welfare recipients, which risked triggering public concerns about the deservingness of recipients (IMF, 2021). The highly publicized new JobKeeper scheme, which ran for a year until 28 March 2021, provided grants to firms to pay workers $750 per week in the first 6 months of the scheme, about 65% of the median wage (ABS, 2020a). From 28 September 2020, this subsidy was reduced to $600 per week for workers who worked over 20 hours per week and $375 for workers who worked less than 20 hours per week, and reduced further to $500 and $325 respectively from 4 January 2021 (FWO, n.d.). Initially, for firms to be eligible for JobKeeper, they had to anticipate a 30% reduction in annual turnover (50% for businesses with turnover over $1 billion). From 28 September 2020, this was assessed against actual business turnover in the third quarter of 2020, leading to a reduction in firms receiving the support (Borland and Hunt, Reference Borland and Hunt2023). As in the UK, these wage subsidies were more generous than standard unemployment benefits but visibly linked to employment status.
Another significant design feature of the wage subsidies was that some recent immigrants and temporary residents were excluded. In the UK, subsidies were restricted to citizens, permanent residents, and temporary visa holders with working rights. In Australia, even temporary visa holders, asylum seekers on bridging visas, and New Zealanders on special category visas were excluded.Footnote 1 While some viewed these policies as harsh and ungenerous, they signaled to others that costly wage subsidies were restricted to those with strong claims to membership in national society. Welfare chauvinism was thus a design feature, aligned with research findings that support for redistribution is lower when it is perceived to benefit out-groups (e.g., Alesina and Glaeser, Reference Alesina and Glaeser2004; Alesina and Stantcheva, Reference Alesina and Stantcheva2020; Alesina et al., Reference Alesina, Baqir and Easterly1999b; Avdagic and Savage, Reference Avdagic and Savage2024; Careja and Harris, Reference Careja and Harris2022; Magni, Reference Magni2021; Page and Shapiro, Reference Page and Shapiro1983; Rueda, Reference Rueda2018).
Fairness beliefs and policy designs were less likely to bolster public support for other kinds of policy intervention. Many people are attracted to narratives that organized economic interests capture policymaking and extract rents from government programs, including in Australia and the UK (Culpepper et al., Reference Culpepper, Shandler, Jung and Lee2025). Thus, programs that benefit these special interests are more likely to be seen as facilitating free riding. Assistance to large firms may also be judged as unnecessary if these firms could manage by drawing on reserves and credit lines, and by reducing profits and dividend payments. In contrast, people could view assistance to small firms as more justified because they tend to be less politically influential and more vulnerable to collapse.
Deviations from fairness norms are especially likely to apply to financial stabilization measures. These are easily perceived as benefiting large, sophisticated, and politically influential financial institutions. Memories of unpopular bank bailouts during the GFC could have resurfaced beliefs about financial-sector rent seeking.Footnote 2 Central bank interventions in key financial markets from March 2020 – many of which reprised GFC schemes – generated outsized gains for already wealthy households and the owners and managers of large banks and firms.Footnote 3
We expect that many people will also have concerns that specific forms of household financial assistance risk breaching fairness norms. Debt relief in particular is likely to trigger reciprocity concerns among people who see themselves as having worked hard to pay off their own debts (J. Ahlquist et al., Reference Ahlquist, Copelovitch and Walter2020). Another strong social norm, that debts should be repaid, is likely to temper support for the collective modification of debt contracts (Graeber, Reference Graeber2011; Guiso et al., Reference Guiso, Sapienza and Zingales2013. In contrast to employees experiencing a loss of income in the pandemic, there are few reasons to expect solidarity between individuals with different varieties of debt (e.g., mortgage, credit card, and student debt). Many people may worry less about the costs that debt relief imposes on banks than the possibility that it produces windfall wealth gains for leveraged beneficiaries – a common objection during the GFC (Matthijs and McNamara, Reference Matthijs and McNamara2015; Skocpol and Williamson, Reference Skocpol and Williamson2016).
Judging which borrowers merit debt relief is especially challenging for others because information asymmetries can produce adverse selection (Akerlof, Reference Akerlof1970). In a deep crisis, some borrowers may already be benefiting from interest rate reductions. Other people may also doubt why individuals already obtaining generous wage subsidies should also need debt relief to maintain debt repayments. The same consideration is likely to apply to payment holidays for renters. Rent relief also imposes costs on private landlords, many of whom are now in the middle class.
Policy design considerations likely compounded these concerns. Generally, financial interventions are less easily understood by the public. The conditionalities attached to them can be difficult even for the well informed to interpret. The website summary provided by the Bank of England (https://www.bankofengland.co.uk/coronavirus) was in much plainer language than that provided by the Reserve Bank of Australia (https://www.rba.gov.au/covid-19/), but central banks have struggled to close the communications gap with the mass public (Braun, Reference Braun2016). Many people may lack trust in the capacity of governments and financial institutions to overcome such informational asymmetries to deliver debt relief to the truly deserving.
In short, debt relief and financial market interventions are more opaque and raise more fairness concerns than wage subsidies, which some could judge to obviate the need for additional debt or rent relief. Thus, we expect relatively muted public support for financial interventions.
Personal pandemic exposure
An alternative theory is that people’s policy preferences are shaped by their direct experiences of the pandemic. The novelty of the pandemic and its unusual severity might have reduced the importance of individuals’ existing fairness beliefs and heightened the impact of variations in direct personal experience. Specifically, individuals’ personal losses and experiences – and those of their close family and friends – could loom larger in their preferences.
Yet the direction of this effect is unclear. One possibility is that close personal experience with the health and economic consequences of COVID-19 or with lockdowns heightened perceptions that losses were beyond anyone’s control, generating more support for economic interventions (Ferwerda et al., Reference Ferwerda, Magni, Hooghe and Marks2024; Harell et al., Reference Harell, Banting, Kymlicka and Wallace2022; Haverland et al., Reference Haverland, van der Veer and Onderco2022; van Oorschot, Reference van Oorschot2006). A different possibility is that close personal experience with health and economic loss could heighten perceptions of competition for scarce public resources, increasing selfishness and concerns about policies benefiting others (Ferwerda et al., Reference Ferwerda, Magni, Hooghe and Marks2024; Haverland et al., Reference Haverland, van der Veer and Onderco2022). Those people who experienced extended lockdowns and who viewed them as a heavy-handed policy choice could have exhibited less support for public assistance.
Partisanship and perceptions of inequality
Patterns of support for economic policy interventions could also vary by partisanship and associated perceptions of inequality. Partisan polarization during the pandemic was visible on issues like the danger that COVID-19 posed to humans, the efficacy and risks of vaccination and mask-wearing, and the net benefits of lockdowns and school closures (Druckman et al., Reference Druckman, Klar, Krupnikov, Levendusky and Ryan2021; Gadarian et al., Reference Gadarian, Goodman and Pepinsky2022; Macedo and Lee, Reference Macedo and Lee2025. Elites and influencers politicized these issues, keeping partisan divides salient (Adida et al., Reference Adida, Dionne and Platas2020).
Although fairness norms tend to be widely endorsed, they are not uniformly distributed. Left-wing partisans tend to value equality and reciprocity more than right-wing partisans, who value proportionality more strongly and who are more likely to believe that market inequality derives from variations in effort and merit (Atari et al., Reference Atari, Haidt, Graham, Koleva, Stevens and Dehghani2023, p. 1180).Footnote 4 Indeed, other research finds that, in the United States, partisan attachments shaped attitudes toward public interventions during the pandemic (Rodriguez et al., Reference Rodriguez, Gadarian, Goodman and Pepinsky2022). Generally, different economic interventions during the pandemic could have triggered concerns about their potential to benefit undeserving groups, activating attachments to partisan narratives.
We expect partisan differences to feature most significantly in case of patterns of support for wage subsidies because social insurance schemes are more prominent in political party conflict. This makes them more likely to activate ideological differences than less familiar financial interventions. In addition, as noted earlier, modification of household liabilities and central bank interventions in favor of large firms and banks are likely to generate concerns among partisans on both sides.
Other factors
Many other factors could shape individual preferences on economic intervention, which we consider as control variables. Notably, policy preferences might divide along material cleavages. A large literature argues that support for redistribution and social insurance decreases as income and education/skills rise and as risk exposure falls (Alesina and La Ferrara, Reference Alesina and La Ferrara2005; Ansell and Samuels, Reference Ansell and Samuels2014; Gingrich and Häusermann, 2015; A. S. Jensen and Wiedemann, Reference Jensen and Wiedemann2023; Rehm, Reference Rehm2009, Reference Rehm2011, Reference Rehm2016; Rehm et al., Reference Rehm, Hacker and Schlesinger2012). Higher wealth and financialization (e.g., financial asset ownership and credit access) can erode support for social insurance, although more leveraged asset holders could exhibit stronger support for wage subsidies, central bank support schemes, and payment holidays to maintain their consumption and asset positions (Ansell and Samuels, Reference Ansell and Samuels2014; Hariri et al., Reference Hariri, Jensen and Lassen2020; Margalit and Shayo, Reference Margalit and Shayo2021).
Very non-financialized people, or the ‘underbanked’ – those lacking formal employment and access to mainstream financial services – may be a special case. Some studies find that individuals with low credit access are more supportive of public income insurance (J. S. Ahlquist and Ansell, Reference Ahlquist and Ansell2017; A. S. Jensen and Wiedemann, Reference Jensen and Wiedemann2023; Markgraf and Rosas, Reference Markgraf and Rosas2024; Wiedemann, Reference Wiedemann2021). However, pandemic wage subsidies paid via employers (as in Australia and the UK), debt relief, and central bank support schemes would not benefit most underbanked people.
Since many studies find that ethic minorities are relatively supportive of redistribution and social insurance (Alesina and Glaeser, Reference Alesina and Glaeser2004; N. Finney et al., Reference Finney, Nazroo, B́ecares, Kapadia and Shlomo2023; Gray et al., Reference Gray, Hunter and Lohoar2011), we also consider whether preferences differ by ethnicity.
Finally, inter-generational differences could play a role. The younger and middle-aged are more vulnerable than older citizens to employment income losses and, because they are typically more leveraged, to asset price falls and credit market disruption. This could make them more supportive of wage subsidies, payment contract relief, and central bank interventions than older citizens, who tend to prioritize low inflation and pension benefit spending (Vlandas, Reference Vlandas2018, Reference Vlandas2023). Older citizens tend to have more assets and less leverage, which could lessen their support for public wage subsidies and interventions to support credit flows and household debt relief. They are also more dependent on income streams from housing, state pensions,Footnote 5 and private pensions. Since the latter are often defined benefit schemes, older citizens could be less vulnerable than others to adverse stock and bond market movements (Bojar and Vlandas, Reference Bojar and Vlandas2021).
Hypotheses
Our primary argument is that people’s preferences about economic policy interventions during the pandemic will be shaped by fairness beliefs and associated policy design, and that this will generate more support for wage subsidies than for other economic interventions. We also assess alternative hypotheses that individuals’ direct personal experience of the pandemic, their partisanship, and their perceptions of inequality shaped policy preferences.
Primary hypotheses: Fairness beliefs and policy design
H1a: Citizens are likely to be more supportive of wage subsidies than other economic policy interventions during the pandemic.
H1b: Citizens are likely to be more supportive of central bank assistance to households and small firms than to large firms and banks.
Alternative hypotheses: personal pandemic experience, partisanship, and perceptions of inequality
Since the direction of impact of personal experience on policy preferences in the pandemic is uncertain, our alternative hypotheses are:
H2a: Citizens who directly experience the negative health or economic consequences of the pandemic, or who have a more prolonged direct experience with lockdowns, are likely to be more supportive of all economic policy interventions.
H2b: Citizens who directly experience the negative health or economic consequences of the pandemic, or who have a more prolonged direct experience with lockdowns, are likely to be less supportive of all economic policy interventions.
H3: Left-wing partisans and citizens who perceive inequality is high and increasing will be more supportive of wage subsidies than of other economic policy interventions during the pandemic.
Data
We worked with Qualtrics to conduct our survey in Australia and the UK from February to May 2021.Footnote 6
We restricted our sample to respondents eligible to vote (age 18 or older), with crossed quotas for age, gender, ethnicity, and location (state/region) to obtain a nationally-representative sample. We oversampled from two additional strata: (1) ethnic minorities (Black, Asian, and minority ethnic, or ‘BAME’ in the UK, and Aboriginal and Torres Strait Islander, or ‘ATSI’ in Australia), and (2) ‘underbanked’ individuals.Footnote 7 The sample size in the UK consists of a total of 2,728 adults, of which 789 reported as BAME, and 308 as underbanked. The Australia sample consists of 2,627 adults in total, of which 204 reported as ATSI and 304 as underbanked.
Cases
We selected Australia and the UK because of their very different pandemic experiences and their sharing of other important characteristics. In the first year of the pandemic, as Figure 1 indicates, Australia was far less affected than the UK by the pandemic. Infection and death rates were far lower in Australia. At the time of our survey, the UK had just emerged from the peak wave of COVID infections and deaths. By the mid-point of our survey in late March 2021, the UK had the highest measured total COVID-19 death rate per million among G20 democracies, whereas Australia had the lowest. This sharp difference provides us with particular leverage on the impact of pandemic experience on respondent preferences. Australia’s bans on international and often domestic travel ensured that large parts of the country were relatively unaffected and local lockdowns were sometimes avoided entirely, with the notable exception of Melbourne, the capital city of the state of Victoria. This strategy was very different from the UK’s, which we account for in our empirical strategy by splitting our sample.
Daily new confirmed COVID-19 infections and deaths, per million people: Australia and UK, January 2020–July 2021.
Source: Mathieu et al. (Reference Mathieu, Ritchie, Rod´es-Guirao, Appel, Giattino, Hasell, Macdonald, Dattani, Beltekian, Ortiz-Ospina and Roser2020)
Despite its failures in preventing infections and deaths, the UK had one of the world’s fastest vaccine rollouts, significantly outpacing Australia’s (Figure 2). The UK had high vaccination rates for both the first dose and the second dose (achieving ‘full vaccination’), whereas vaccination was relatively delayed in Australia. If our alternative hypotheses that pandemic experiences mattered for public support for economic policy interventions are valid, we would expect to see significant divergence between Australia and the UK.
People receiving first vaccine dose and fully vaccinated, per 100 people: Australia and UK, January 2020–July 2021.
Source: Mathieu et al. (Reference Mathieu, Ritchie, Rod´es-Guirao, Appel, Giattino, Hasell, Macdonald, Dattani, Beltekian, Ortiz-Ospina and Roser2020)
Australia and the UK also have many political and cultural similarities. Center-right governments with populist tendencies were in office in both countries during the pandemic. Both countries are prime examples of liberal welfare state systems. Redistribution is lower in Australia and the UK compared to continental European peers and more conditional/means-tested than in social democratic systems. Our selection thus creates a harder test for H1a, since individual support for generous wage subsidies in these countries may be grounded in more egoistic rationales than in many European democracies.
Despite their political orientations, governments in both countries intervened extensively during the pandemic, including by providing the unusually generous wage subsidies described above. In the UK, support measures began in March 2020 following the implementation of a strict national lockdown. Both governments also undertook financial interventions aimed at stabilizing household, corporate, and financial sector balance sheets. Their payment holiday schemes were broadly similar. In the UK, the Financial Conduct Authority (FCA) issued guidance to lenders in March 2020 to allow mortgage holders a three-month payment holiday, which was extended to six months in November 2020. Parallel consumer credit measures were also introduced for loans and credit cards. All payment holidays had to end by July 2021, with applications required by the end of March 2021. In total, nearly 3 million mortgage payments were deferred (Cromarty, Reference Cromarty2023).
The UK government introduced no equivalent measures benefiting household renters, though it provided some protection against eviction. In Australia, the Prudential Regulation Authority (APRA) granted regulatory concessions allowing banks to offer mortgage holidays from March 2020. Most banks allowed mortgage holders to defer payments for up to six months. In July, APRA announced that banks could extend payment holidays to 10 months, with all deferrals ending by the end of March 2021. Some banks also offered relief on credit card payments, though the duration and terms of these deferrals varied from bank to bank. Some relief for household renters was provided by Australian state governments. In the two largest states (New South Wales and Victoria), renters were given temporarily greater protection from eviction by landlords. Some states also provided modest subsidies to renters and landlords.
Central banks in both countries introduced comprehensive packages of monetary easing that included interest rate cuts, purchases of government bonds, and, in the case of the UK, bonds of non-financial firms. Additional liquidity support was provided to the financial system but stopped short of purchasing assets directly from banks (Bank for International Settlements, 2020; Cantú et al., 2021). No central bank support measures were targeted directly at households.
Central bank support for firms was more targeted at small and medium enterprises (SMEs) in Australia.Footnote 8 In March 2020, the Reserve Bank of Australia introduced a new low-cost A$90 billion funding facility for banks to lend to SMEs. This facility was subsequently expanded to A$200 billion at an even lower rate, with access extended to June 2021 (RBA, n.d.). Similarly, in the UK, the Bank of England introduced a low-cost funding facility to banks, lending more to SMEs. With the Treasury, the Bank of England also purchased short-term debt from larger firms to support their liquidity. Also with the Treasury, the Bank of England rolled out three loan guarantee schemes targeted at both SMEs and larger firms. A total of £352bn of liquidity and loan guarantees was made available to businesses (IMF, 2021).
Our survey
We used a conjoint experiment to study Australian and UK citizens’ preferences for different COVID-related economic policy interventions. These interventions included the level and target of income support policies, debt relief, and central bank support schemes. A conjoint design approach is well suited to our study, as it permits ‘researchers to estimate the causal effects of multiple treatment components and assess several causal hypotheses simultaneously’ (Hainmueller et al., Reference Hainmueller, Hopkins and Yamamoto2014, p. 1).
Our aim was to explore preferences for different policy interventions while taking into account the scope and beneficiary of a given intervention. Each respondent was asked to evaluate five separate pairs of COVID-related economic policy interventions (‘attributes’) with randomly generated levels that varied across the two sets. After each pair of policy packages, a forced-choice question asked respondents to select which of the two policy packages they preferred. For each policy intervention pair that a respondent viewed, depending on the intervention (policy attribute), the scope, duration, and beneficiary (policy level) were randomly assigned.
Table 1 shows the attributes and levels we used. The attributes include the level and duration of wage subsidies, the target and duration of payment holidays, and the target of central bank support schemes.Footnote 9
List of attributes and attribute levels in the conjoint experiment
These attributes are based on the key fiscal and monetary policy interventions implemented in Australia and the UK following the outbreak of COVID-19. The set of levels for each policy intervention was chosen based on pretesting results and actual policy experience. Given the complex package of measures we asked respondents to assess, our conjoint design presented the treatments in an invariant order to reduce cognitive overload. While consistent with some other studies (Carnes and Lupu, Reference Carnes and Lupu2016; Chilton et al., Reference Chilton, Milner and Tingley2020; Hainmueller et al., Reference Hainmueller, Hangartner and Yamamoto2015; Huff and Kertzer, Reference Huff and Kertzer2018), this assumes that attribute order does not affect our results. We are mindful that this possible bias may limit our ability to compare the effects of attributes.Footnote 10
We arrange our data so that the policy package
$k$
of binary comparison
$j$
to respondent
$i$
represents a unique observation. Each respondent completed five iterations of the experimental binary comparison, resulting in ten observations per respondent and 53,550 observations overall.
We then use ordinary least squares regression analysis to relate our outcome – whether the respondent selected a particular policy package – to binary indicators capturing the randomly assigned presence of a policy level (for instance, a wage subsidy of 80 percent of the pre-crisis wage). Our models take the following form:
$$Y = {\alpha _0} + \beta {X_{ijk}} + {\varepsilon _{ijk}}$$
where the dependent variable
$Y$
refers to the policy package choice for any given respondent
$i$
presented with policy package
$k$
in binary choice comparison
$j$
;
$X$
represents a vector with the attributes of the policy package presented to the respondent, and
$e$
denotes the error term. We cluster our standard errors by respondents because each respondent evaluated ten policy packages.
Conjoint results
The results from the pooled sample are shown in Figure 3. The left panel reports estimates of the average marginal component effect (AMCE), which capture the elicited respondent preferences from the randomly generated combinations of levels for each attribute (Hainmueller et al., Reference Hainmueller, Hopkins and Yamamoto2014). The AMCEs represent the average influence of the different policy levels on the probability that a particular policy is preferred by a respondent. The interpretation of each estimate is relative to that attribute’s reference category shown in Table 1. The right panel reports estimates of marginal means (MMs), which capture the overall favorability of respondents toward particular policy packages with a particular level, irrespective of all other levels. In forced-choice conjoint designs, MMs have a straightforward interpretation as probabilities (Leeper et al., Reference Leeper, Hobolt and Tilley2020).
AMCEs and MMs from conjoint survey experiment, pooled.
The results indicate overall support for economic interventions during the pandemic. Notably, the MMs show that for each policy attribute, we observe opposition to nonintervention. However, consistent with H1a, we find the strongest support for generous, extended wage subsidies. Compared to the baseline of no wage subsidies, respondents favor subsidies of 50–80 percent of average pre-pandemic wage levels, with a preference for higher levels. Respondents also favor subsidies to be paid for at least 12 months.
We found significantly less strong support for contractual payment holidays. Respondents prefer that debt relief is provided as blanket payment holidays. Interestingly, in contrast to other payment holidays, relief restricted to mortgage holders and renters is less popular. We interpret this to suggest that respondents favor holidays for mortgage payments but less so for rental contracts, unless the borrower is facing COVID-related distress or the relief extends to consumer and unsecured debt. As with wage subsidies, respondents were slightly more supportive of payment holidays of at least 12 months’ duration. Consistent with H1b, we also find respondents are highly favorable toward central bank support for households and small firms, and more opposed to support for large firms and banks.
Overall, these pooled results provide tentative support for H1a and H1b. They also reveal variation in patterns of support toward social insurance of employment/income risk on the one hand, and, on the other hand, financial interventions that stabilized household, corporate, and financial sector balance sheets.
This variation is consistent with our argument that people apply fairness considerations differently depending on the type and design of policy interventions. Generous, extended wage subsidies in Australia and the UK were very popular despite their deviation from prior policies. Standard concerns about the risk that free riders would benefit seem to have been lessened by the economic shock’s viral origins (Cavaillé, Reference Cavaillé2023, p.21). In the Australian and British cases, policy design was probably also important in minimizing such concerns. Wage subsidies were universally provided to citizens, they were capped so as not to disproportionately benefit very high earners, and they were paid to employers rather than directly to employees. However, other payments to firms are less popular, probably because they raised more concerns about their potential to benefit less needy borrowers.
Respondents’ greater skepticism toward central bank interventions supporting large firms and banks also points to the significance of reciprocity considerations. Narratives that large, politically influential private sector actors would benefit disproportionately from the COVID interventions were aired in critical media coverage at the time of our survey (e.g., Herald, Reference Herald2021; Times, Reference Times2020. A high-profile corruption scandal broke in both countries in March 2021, in which former UK Prime Minister David Cameron was revealed to have lobbied his own government to provide COVID emergency loans to Greensill Capital, a failed finance company run by a high-profile Australian CEO to whom Cameron was a paid advisor (Bourke, Reference Bourke2021; Times, Reference Times2021). Footnote 11
As a robustness check, we analyzed the conjoint experiment by country. Appendix Figure A.8 reports the marginal means of each attribute level by country. Overall, the results demonstrate parallel developments and are consistent across countries, which increases our confidence in the robustness of the findings.
Pandemic experiences and partisanship
We use heterogeneous treatment effects (HTEs) to consider alternative explanations. Details about the operationalization of all our variables and summary statistics are shown in Appendix A.1.
Pandemic experience
We first consider whether people’s pandemic experiences shaped their policy preferences. We begin by exploring whether individual attitudes vary as a function of their direct experience with pandemic-specific economic or health loss.Footnote 12 We examine whether individuals experienced a personal economic or health loss in their household and in their network of friends and family. Contrary to H2a and H2b, Figure 4 shows no significant differences in policy preferences related to direct personal experience with pandemic-specific economic or health losses.
MMs from conjoint survey experiment, by economic and health losses, pooled.
Our data also enable us to consider whether individual preferences relate to pandemic-specific experiences in their local environment. We examine respondents’ environments by looking at the local intensity of the pandemic, as measured by three different indicators: infections, deaths, and the pace of the vaccine rollout.Footnote 13 We geo-locate our respondents at the lowest level of aggregation possible based on data availability (the state level in Australia, and the local authority district (LAD) in the UK). Separately for each indicator (infections, deaths, and vaccinations), we group and compare respondents residing in areas where local conditions in the prior three months were above the country-specific median with those respondents residing in areas where local conditions were below the country-specific median. However, again contrary to H2a and H2b, as shown in Figure 5, we find no significant differences related to pandemic context.Footnote 14
MMs from conjoint survey experiment, by infection rate, death rate, and vaccine rate in the 3 months prior, pooled.
We also consider whether lockdown experiences shape patterns of support. The varying use and intensity of lockdowns in the two countries raise complications. Whereas during the time of our survey, lockdown restrictions were eased for our UK sample, virtually all Australian respondents experienced no domestic restrictions, though stringent international travel bans remained in place.Footnote 15 This prevents us from using the pooled sample to examine HTEs for lockdown restrictions in a meaningful way. By splitting the sample, we can also address the possibility that the different lockdown and travel policies adopted by both governments could make them less comparable in this area.
In the UK, as outlined in Appendix A.2, lockdown restrictions were eased over a period of several months following a road map announced in February 2021. As our survey was in the field during this time, we can exploit how variation in intensity of UK lockdown restrictions shapes patterns of support for economic policy interventions.Footnote 16 Our estimation strategy splits the UK sample into two groups, which consist of respondents who took the survey before schools were reopened on 8 March 2021 (full lockdown) and those who took the survey from 8 March 2021 (after the first step in the lockdown easing roadmap).Footnote 17
In Australia, using postcode-level data, we focus on the cumulative number of days under lockdown each respondent had experienced when they took the survey.Footnote 18 Most respondents avoided lockdowns, but a subset of respondents in Victoria experienced lockdowns for a week or more, with residents of Melbourne exclusively experiencing the most prolonged use of these restrictions, stretching for 100 days or more. We thus focus on those Melbourne residents in our sample who experienced the most prolonged use of lockdown restrictions. Our estimation strategy splits the Australia sample into two groups, consisting of Melbourne-based respondents who experienced lockdown for 100 days or more and all other respondents.Footnote 19
Using these country measures to address variations in patterns of support due to differing experiences with lockdown, we find no significant differences across the different subgroups (See Figure 6). This is once again contrary to H2a and H2b. One possible reason for these null results is that personal experiences of lockdowns could have triggered empathetic responses in the early stages of the pandemic, as Herrera et al. (Reference Herrera, Ordoñez, Konradt and Trebesch2020) find. However, by the time our survey was conducted, more ‘lockdown fatigue’ could have set in for more citizens, raising the political costs of continuing them. These divergent reactions could have offset each other, contributing to the null result.
MMs from conjoint survey experiment, by lockdown experience and residence in Victoria (Australia), and survey date (UK).
Overall, and in contrast to other studies, we find no evidence that patterns of support relate to the pandemic context – whether through direct personal exposure, the local environment, or lockdown measures. Moreover, we find consistent results in the conjoint experiment across the two countries despite their large differences in the intensity of the pandemic and lockdown experiences at the time of our survey.
One possible explanation is that other studies focus more explicitly than ours on the redistributive character of government interventions and highlight their potential availability to out-groups, including citizens of foreign countries (Ferwerda et al., Reference Ferwerda, Magni, Hooghe and Marks2024, p. 17; Haverland et al., Reference Haverland, van der Veer and Onderco2022, p. 687). Such measures may tap more strongly into respondents’ welfare chauvinism. Our results are more in line with those of Bremer et al. (Reference Bremer, Kuhn, Meijers and Nicoli2023, p. 19), who find in a five-country study of European attitudes toward fiscal transfers that exposure to the economic risks of the pandemic was unrelated to differences in support for the purpose and distribution of the transfers.
Partisanship and inequality perceptions
We find more support for H3. The left panel of Figure 7 indicates that there are significant differences in policy preferences by left-right orientation. Left-wing respondents tend to support more extended wage subsidies than right-wing respondents, suggesting that enduring partisan beliefs shaped preferences through the pandemic.
MMs from conjoint survey experiment, by partisanship, perceived society type, and perceived inequality trends, pooled.
Partisan divides are linked less strongly with central bank support schemes benefiting large firms and banks. Our interpretation of these results is that although central bank interventions can be highly redistributive, they have been less prominent in political party conflict than social insurance schemes, which more readily activate partisan differences. As noted earlier, central bank interventions can trigger concerns among both left and right partisans that they will generate outsized market gains for business and the wealthy, consistent with a narrative that the economy is rigged to benefit elites (Culpepper et al., Reference Culpepper, Shandler, Jung and Lee2025). Many people in Australia and the UK were already concerned about existing levels of inequality (Benson et al., Reference Benson, Duffy, Hesketh and Hewlett2021; Duffy et al., Reference Duffy, Murkin, Hewlett, Benson and Hesketh2021). As noted earlier, central bank interventions in key financial asset markets generated outsized gains for already wealthy households and the owners and managers of large banks and firms.
We used the ‘perceived society type’ approach to gauge individual perceptions of the level of inequality. Drawing on a question from International Social Survey Project, respondents were shown five diagrams and verbal descriptions of different types of society, labeled Type A, B, C, D, and E, and asked which best described their country.Footnote 20 We use estimates from Gimpelson and Treisman Gimpelson and Treisman (Reference Gimpelson and Treisman2018) to calculate the subjective Gini coefficients of each diagram, with Type A being by far the most unequal (0.42), followed by Type B (0.35) and Type C (0.30). The Gini coefficients for Type D (0.20) and Type E (0.21) are similar, and, following Gimpelson and Treisman Gimpelson and Treisman (Reference Gimpelson and Treisman2018), we combine responses for these two types into one category.
The center panel of Figure 7 shows that those respondents who perceive a more extreme level of inequality in society (Type A) are more supportive of generous wage subsidies than those who perceive a more equal society (Types B, C, and D/E).Footnote 21 We find similar results in the right panel of Figure 7 among those who perceive the level of inequality as rising.Footnote 22
The existence of these differences, rooted in partisanship and inequality perceptions, is consistent with H3 and aligns with other findings that ideological divides endure in the face of shocks that are perceived as shared and as beyond individual control. Ferwerda et al. Ferwerda et al. (Reference Ferwerda, Magni, Hooghe and Marks2024), for example, find that left-wing and right-wing respondents display varying discrimination toward outgroups when considering public health guidelines and vaccine prioritization. Rovny et al. (Reference Rovny, Bakker, Hooghe, Jolly, Marks, Polk, Steenbergen and Vachudova2022) also find that political parties interpreted the crisis through partisan lenses. Partisanship also shaped how voters evaluated incumbent party performance in managing the pandemic (Druckman et al., Reference Druckman, Klar, Krupnikov, Levendusky and Ryan2021; Ferwerda et al., Reference Ferwerda, Magni, Hooghe and Marks2024; Flores et al., Reference Flores, Cole, Dickert, Eom, Jiga-Boy, Kogut, Loria, Mayorga, Pedersen, Pereira, Rubaltelli, Sherman, Slovic, V¨astfja¨ll and Van Boven2022; Gadarian et al., Reference Gadarian, Goodman and Pepinsky2022; Gollwitzer et al., Reference Gollwitzer, Martel, Brady, Pärnamets, Freedman, Knowles and Van Bavel2020). Our analysis suggests that attitudes to economic policy interventions were also a key, yet understudied, aspect of partisan divides during the pandemic.
Other factors
We considered a range of other social and economic factors that could drive patterns of support for policy interventions. We first addressed divides associated with income, education, ethnicity, asset holdings, wealth, financialization, and economic risk exposure that are commonly found to shape attitudes toward redistribution and social insurance. Yet we find few sizeable and statistically significant differences across these divides. Individuals across the income distribution and across education levels show similar patterns of favorability toward economic policy interventions (Appendix Figure A.16 and A.17). When we consider ethnic divides we also find little variation within each country and in the pooled sample (see Appendix Figure A.18). Divides between homeowners and renters are also not statistically apparent in our results, even when we consider the potential mitigating impact of house price levels and appreciation (Appendix Figure A.19, A.20, A.21, A.22, A.23, A.24, A.25). Likewise, we fail to uncover any statistically meaningful differences among those with higher and lower levels of financial asset ownership (Appendix Figure A.26). It is notable that we find no evidence that variation in respondents’ material position in asset markets shapes support for central bank policies benefiting large firms and banks. This supports our interpretation that fairness concerns about the disproportionate market gains arising from these interventions dominated material self-interest.
We do, however, find that underbanked status and economic risk exposure are related to different patterns of support. Figure 8 shows that underbanked respondents are unusually indifferent to debt market and central bank interventions. This could be because these respondents have less understanding of and experience with these policies, or because they perceive little direct benefit from them. Consistent with the latter interpretation, Figure 8 also shows that recipients of the CJRS or Jobkeeper schemes (‘furloughed’ workers) are more favorable than the general public toward generous versions of these schemes. Interestingly, however, we find no significant differences related to employment status. Underbanked respondent support for wage subsidies also seems inconsistent with a materialist interpretation, although it underscores the broad support for social insurance of employment/income risk.
MMs from conjoint survey experiment, by financial inclusion, furlough, and employment status, pooled.
We also find that older citizens are less favorable toward generous and costly wage subsidies (see Appendix Figure A.29). They are also less favorable toward relief for renters, unless they are facing COVID-related distress or such measures are extended to consumer and unsecured debt. This could reflect older citizens’ lesser dependence on employment income and the higher priority they tend to place on low inflation.Footnote 23 Older citizens are also more strongly opposed to central bank support schemes for large firms and banks, which could reflect their lesser vulnerability to stock market movements. These findings imply that older respondents were less pro-social during the COVID crisis than is suggested by some research. However, older respondents are more favorable toward central bank support for households and small firms, suggestive of their willingness to support some policies that mitigate the general hardship associated with lockdowns and other protective health measures.
Conclusion
We have shown that public support for the economic policy interventions undertaken during the COVID crisis in Australia and the UK was generally positive, but significantly higher for wage subsidies than for financial interventions. We found little support for one possible explanation for this finding, which focuses on the impact of direct pandemic experiences on people’s policy preferences. Nor were a range of socio-economic divides that have often been found to influence attitudes toward redistribution and social insurance important. This included factors such as income, education, ethnicity, wealth, and financialization.
Instead, our findings suggest that policy preferences were more strongly related to fairness beliefs, policy design, and partisan identities. Support for generous wage subsidy programs, while still varying on partisan lines, was more strongly aligned with intuitive fairness beliefs. The viral nature of the crisis and similar policy design features in the wage subsidy programs in both countries probably minimized concerns about their redistributive impact and their potential to facilitate free riding and opportunism. Conservative governments in Australia and the UK retained traditional conditionalities for these subsidies, including requiring strong evidence of pandemic-related loss of employment income, which likely further reassured people inclined to skepticism about social insurance.
Were most of our respondents fearful of the personal economic risks they faced in the pandemic? In response to negative economic shocks, self-interested motivations can lead individuals to be more supportive of social insurance and redistribution, though the measured effect of perceived risk of loss rather than actual loss is more mixed (Ansell, Reference Ansell2023, Margalit, Reference Margalit2019, pp. 281–283). COVID generated widespread actual and threatened losses for normally secure groups that face limited risk even in deep financial crises. The acute threat posed by the pandemic to middle-class wage income was unusual, which might explain some of the support for high levels of wage insurance, particularly among our furloughed respondents who had strong self-interest motivations. Yet we found that personal experience with actual health and economic losses during the pandemic was not significant driver of attitudes toward the economic policy interventions. We view this as consistent with our interpretation that there was a general understanding that wage insurance was justified largely on fairness grounds.
Lower support for financial market interventions, including debt relief, rent subsidies, and central bank interventions benefiting large firms and banks, was more likely to trigger concerns about undeserved gains. Upward redistribution to politically influential and potentially opportunistic large firms and banks in the COVID-19 crisis, as in the GFC, was clearly less popular (Bermeo and Pontusson, Reference Bermeo and Pontusson2012; Chwieroth and Walter, Reference Chwieroth and Walter2019; Culpepper et al., Reference Culpepper, Shandler, Jung and Lee2025; Frieden and Walter, Reference Frieden and Walter2017; Gomez, Reference Gomez2024; McCarty et al., Reference McCarty, Poole and Rosenthal2013).
Respondents also discriminated between the benefits of debt relief, rental payment relief, and central bank interventions to households and SMEs on the one hand, and interventions supporting large firms and banks on the other. Support was noticeably stronger for the former category of interventions than the latter. Corporations, especially large ones, are composite, impersonal institutions with whom individuals empathize less easily. This may be especially true in an era in which many large businesses have been highly profitable and their senior managers exceptionally remunerated (Piketty, Reference Piketty and Goldhammer2014, p. 334).
Fairness considerations were, however, not the only important factor. Beliefs about economic disparities, partisanship, and age significantly shaped attitudes toward economic policy interventions. People who perceived economic disparities to be low, right-wing partisans, and more elderly respondents were less supportive of income, debt, and asset market interventions than their progressive and younger counterparts. A more notable finding, especially because we found major asset and financialization divides were generally unimportant, was that very financially marginalized respondents were also less supportive of debt market and central bank interventions. These ‘underbanked’ respondents either had less understanding of these policies or were skeptical about their benefits. More research is needed to investigate the preferences and motivations of this less understood but significant group: nearly one in five adults in Australia and one in seven adults in the UK are financially excluded, unable to access safe, affordable, and fair financial services (Godinho, Reference Godinho2017; LexisNexis Risk Solutions, 2021).
Did our respondents perceive low personal costs from these interventions? This is possible but, we think, unlikely. The heavy fiscal costs of the pandemic measures received extensive media attention in both countries and their inflationary consequences were already apparent by the time of our surveys. Indeed, 45 percent of all our respondents expected inflation to be above target over 2021–22 (see Appendix Figure A.4). Our respondents remained supportive of these interventions despite sharply rising fiscal deficits over 2020–2021, which generally lessen the attachment of individuals to fairness considerations favoring social insurance (Cavaillé, Reference Cavaillé2023, pp.16–17).
Will future economic shocks and associated policy interventions be associated with higher levels of political contestation? The generalizability of our findings may be limited because, although Australia and the UK had quite different pandemic experiences, their welfare state models and policy responses were broadly similar. Although we would expect fairness beliefs and partisan identities to be important in shaping citizens’ policy preferences in other countries, this expectation could be tested in future research. One avenue to address this could be policy responses to climate-induced shocks. These share some similar characteristics with the pandemic interventions, being at least in part exogenous to individual countries and often imposing widely shared risks on citizens. Our findings on pandemic preferences suggest that public support for climate ‘adaptation’ policies in response to shocks imposing large costs on society will continue to be stronger than for pre-emptive climate change mitigation, which entails costly investments today to limit diffuse and uncertain future losses.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1475676525100327
Data availability statement
Replication data for this article will be made available as an online appendix.
Author contributions
Jeffrey Chwieroth: Conceptualization-Equal, Data curation-Equal, Formal analysis-Lead, Funding acquisition-Equal, Investigation-Equal, Methodology-Equal, Project administration-Equal, Resources-Equal, Software-Equal, Supervision-Equal, Validation-Equal, Visualization-Equal, Writing - original draft-Equal, Writing - review & editing-Equal. Andrew Walter: Conceptualization-Equal, Data curation-Equal, Formal analysis-Supporting, Funding acquisition-Equal, Investigation-Equal, Methodology-Equal, Project administration-Equal, Resources-Equal, Software-Equal, Supervision-Equal, Validation-Equal, Visualization-Equal, Writing - original draft-Equal, Writing - review & editing-Equal
Funding statement
We gratefully acknowledge funding from the Research and Impact Fund at the London School of Economics (LSE) and support of the Economic and Social Research Council in funding the Systemic Risk Centre at the LSE (grant numbers ES/K002309/1, ES/R009724/1, and ES/Y010612/1).
Competing interests
None.
Open access
