Economic inequality is a major global burden with perpetually negative individual and population consequences (Chancel & Piketty, Reference Chancel and Piketty2021). Greater income inequality correlates with lower life expectancy (Chetty et al., Reference Chetty, Stepner, Abraham, Lin, Scuderi, Turner and Cutler2016b), suppressed economic growth (Bivens, Reference Bivens2017), and wider political polarization (Voorheis, McCarty, & Shor, Reference Voorheis, McCarty and Shor2015). More than 70% of the global population reside in countries where inequality is rising, exacerbating risks of conflict and slowing economic development (United Nations, 2020).

Policies addressing economic disparities directly through redistributive welfare programs or financial incentives (Barrientos, Reference Barrientos, Midgley, Surender and Alfers (Eds.)2019) have potential to improve population well-being (Thomson et al., Reference Thomson, Igelström, Purba, Shimonovich, Thomson, McCartney and Katikireddi2022). However, most have failed to mitigate growing wealth gaps and need to be integrated with other substantive efforts (Millán, Barham, Macours, Maluccio, & Stampini, Reference Millán, Barham, Macours, Maluccio and Stampini2019). Interventions are typically developed on the perspectives of economists and legislators, which often comprise condensed geographic and socioeconomic viewpoints (Bureau of Labor Statistics, n.d.). Thus, most policies fail to consider true behaviors and challenges of those who have successfully overcome significantly disadvantaged circumstances. Study of these “positive deviants” would better equip policies to support sustained and meaningful upward economic movement (Ruggeri & Folke, Reference Ruggeri and Folke2022).

Reference Chater and LoewensteinChater & Loewenstein provide a valuable opportunity to incorporate this thinking by differentiating systems (s-frame) and individuals (i-frame) in policies. In this commentary, rather than critique or debate that framing, we propose the tremendous potential for impact by incorporating both when designing policies to reduce economic inequality.

Consider how the COVID-19 pandemic added barriers globally to overcoming inequality while disproportionately burdening low-income individuals. The bottom 20% of earners in 2021 were nearly 7% lower than projected before 2020 (Sidik, Reference Sidik2022). Using data from a 60-country study (n = 12,930) on temporal discounting (Ruggeri et al., Reference Ruggeri, Panin, Vdovic, Većkalov, Abdul-Salaam, Achterberg and García-Garzon2022), we classified 12.5% of participants as positive deviants (low-income childhoods yet healthy financial decision makers as adults). Figure 1 highlights how positive deviants were less likely to have been negatively affected economically during the pandemic than the 16.9% that remained low income as adults. Such patterns illustrate the benefits of upward movement (i.e., resilience against crises) and the self-perpetuating harms of economic inequalities (i.e., poverty increases financial vulnerability in a crisis).

Figure 1. Comparison of financial impacts of the COVID-19 pandemic in 2020 between financial circumstances. Each element is ordered by the rate of difference between those experiencing positive/neutral impacts and those experiencing negative impacts by country within each group. Pakistan, Lebanon, and Egypt had only one positive deviant, so the proportion is shaded to avoid skewing perception.

The pandemic catalyzed a proliferation of redistributive initiatives, yet evidence of their substantive effects toward alleviating disparities is mixed. This is not unique to COVID policies: Table 1 summarizes major policies that lacked measurable impact on the trajectory of inequality. This is not a criticism of the policies themselves, but indicates the complexity of the problem and supports reframing policy creation.

Table 1. Examples of i-frame and s-frame policies aimed at reducing income inequality

Despite the minimal impacts of simple redistribution during the pandemic, valuable behaviors were still observed. Figure 2 displays spending patterns for approximately 6,000 low-income individuals that received a US Cares Act Economic Impact Payment of exactly $1,200. For the first $1,200 spent after receiving the stimulus check, 94% overall went to discretionary and daily living expenses. Among that group, about 4% of individuals allocated funds (5% on average) to savings or investments. This trend continued beyond the first $1,200 (3% on average) and is highly consequential: Individuals who saved money will have greater financial well-being over time than individuals who allocated entirely to near-term spending.

Figure 2. Spending patterns in a low-income ($17,240–34,480) group in the United States, split by positive deviants and others, immediate spending post stimulus check and for the first month following the stimulus check.

Finding these distinct patterns among positive deviants presents a potentially meaningful target for financial management policies. Misguided approaches might encourage a low-wage individual to save without addressing the cost of living, thereby superficially effective and simultaneously encouraging high-interest debt like credit cards (Sussman & O'Brien, Reference Sussman and O'Brien2016). Actionable insights on the behaviors of those who have overcome inequality, such as specific saving values (3–5%), may better position policies to reduce economic inequality at the individual level because they factor-in the context.

However, even with effective interventions, individual behavior cannot alone resolve wider structural barriers, such as inequitable pay. Consider that national rates of overcoming inequality are highly varied: In the dataset described earlier, positive deviance rates ranged from 0.8% (Egypt) to 26.2% (Canada). As individuals from countries with the largest income disparities demonstrate greater rates of high-risk behaviors and increased debt accumulation (Payne, Brown-Iannuzzi, & Hannay, Reference Payne, Brown-Iannuzzi and Hannay2017), policies cannot be presumed to be equally effective in all contexts. Heterogeneous national rates of overcoming inequality suggest variability in barriers, resources, and opportunities. For example, greater gender equality is associated with higher rates of female positive deviance (see Fig. 3), with no significant correlation for males. Similar patterns exist in US healthcare, where general wage increases overall were directly associated with decreases in the wage gap (Barry, Reference Barry2021).

Figure 3. Relationship between gender equality and rates of female positive deviance in 59 countries (r = −0.31, P = 0.017). Data from temporal discounting study and the UN's Gender Inequality Index. Gender inequality scores are reversed for easier understanding in the visual. Limitations to the data and analyses can be found in the Supplementary material (Methods and materials).

These patterns demonstrate the need to create policies that also address structural inequality. Such types of policies (see Table S4) have shown promise. Britain's Pay Transparency Initiative (Gamage, Kavetsos, Mallick, & Sevilla, Reference Gamage, Kavetsos, Mallick and Sevilla2020) made a clear, positive impact on reducing wage disparity. Similarly, Dutch requirements for corporate boards to comprise at least one-third women ensure greater participation at a level that removes salary ceilings (Women's Labour Force Participation, n.d.).

Our argument is, therefore, to blend the structural and the individual, rather than limiting to only systems or decision making in isolation. An effective example comes from a World Bank initiative in Uruguay (Ubfal, Reference Ubfal2022), which successfully implemented a work–study intervention to reduce gender economic inequality through equal pay for students. Integrating individual working and financial behaviors along with implementing a policy environment that created equitable work opportunities resulted in significant, positive effects for girls who participated. There were also no negative effects for participating boys.

Despite increased investment into reducing economic inequality, even generous, accessible financial incentives have alone been ineffective at reducing inequality or even modestly advancing the economic position of those receiving payments (Jaroszewicz, Jachimowicz, Hauser, & Jamison, Reference Jaroszewicz, Jachimowicz, Hauser and Jamison2022). Rather than to continue approaches based on assumptions of normative behaviors, policies must concurrently target systemic factors and individual behaviors facing significant social and economic challenges, perhaps by encouraging the choices of peers who have overcome inequality. That means large-scale investment targeting barriers (e.g., pay gaps) must also be met with effective, individually relevant policies that reduce the risk of choices or behaviors that propagate inequality created by challenging circumstances. As demonstrated, such approaches will directly impact the well-being of individuals and populations.