This article is included in the Research on Research, Policy & Culture gateway.
Income inequality (D31), aggregate mark-ups and market power (D43, L13), voucher-based redistribution and in-kind transfers (H23), corporate ESG and social responsibility (M14), panel econometrics and staggered difference-in-differences (C23), innovation and R&D intensity (O32).
This revised version responds to the second reviewer’s methodological and policy concerns by substantially clarifying the empirical interpretation and moderating the strength of the causal claims. We now distinguish more explicitly between full Counterbalance Economics (CBE) as a proposed voucher-based framework and the empirical CBE-adjacent policy indicator used in the paper, which captures real-world rent-recycling mechanisms rather than full implementation of the CBE model. The treatment coding has also been clarified: only China, Malaysia, and Australia provide identifying variation within the 1982–2016 estimation window, while Germany, France, Sweden, and Brazil are discussed as post-sample policy analogues rather than contributors to the main econometric identification.
The methodology and data sections have been revised to address concerns about missing Gini observations, interpolation, and the implausible zero value reported for GDP per capita. We now present the empirical results more cautiously, treating them as evidence from policy analogues rather than as definitive estimates of a fully implemented CBE system. The discussion, limitations, and conclusion have also been expanded to address inflationary transmission, regulatory capture, firm compliance, informality in emerging economies, long-run innovation and growth effects, and the interaction between CBE and existing tax-transfer systems. Overall, the revision strengthens transparency, construct validity, and policy realism while preserving the paper’s core theoretical contribution.
See the authors' detailed response to the review by Marina da Silva Sanches
See the authors' detailed response to the review by D. Tripati Rao and Praveen Kumar
Income inequality has become a central public policy challenge in both developed and developing economies. Empirical research since the mid-2000s links high inequality to slower and less durable growth (Berg & Ostry, 2011; Cingano, 2014) as well as to reduced intergenerational mobility (Chetty et al., 2014) and greater social tensions and health costs (Wilkinson & Pickett, 2009). When income and wealth are concentrated in a few hands, aggregate demand can weaken, and macro-financial fragility may rise (Kumhof & Rancière, 2010), while trust and social cohesion erode (OECD, 2015). Addressing inequality is, therefore, not only a matter of social justice but also of economic prudence: increased disparities undermine productivity, political stability and long-run growth potential (Piketty, 2014). However, despite broad agreement on its importance, achieving a more equitable income distribution has proved difficult. Conventional tools—such as progressive income taxes, cash transfers, and minimum wages—have delivered only limited and sometimes temporary, reductions in inequality (Atkinson, 2015). Political constraints and concerns about efficiency often curb the ambition of these policies, echoing Arthur Okun’s classic “leaky bucket” analogy, in which redistribution may lose efficiency (water) as it moves resources from the rich to the poor.
In the face of these challenges, policy debate has shifted toward market-based mechanisms that enlist private firms as active agents of redistribution. A rich corporate-responsibility literature finds that companies increasingly embrace social objectives—whether to secure reputational capital, satisfy ESG (Environmental, Social, and Governance)-oriented investors, or expand future consumer demand (Bénabou & Tirole, 2010; Eccles, Ioannou, & Serafeim, 2014). Building on this trend, we introduce Counterbalance Economics (CBE), an inter-corporate voucher architecture in which firms issue and accept Countertrade Credits (CCs) that lower-income workers can redeem for goods and services. CCs create a parallel, non-cash income channel that boosts purchasing power at the bottom of the distribution without relying on higher taxes or inflationary finance, echoing the efficiency arguments advanced for well-designed in-kind transfers and complementary currencies (Johnson, 2018; Teasdale & Aiken, 2019). By embedding redistribution within regular commercial exchange, CBE aligns with the “shared-value” paradigm that sees profitable opportunities in meeting social needs (Porter & Kramer, 2011) and with evidence that firms with credible ESG commitments enjoy lower capital costs and stronger long-run performance (Krüger, 2015; Liang & Renneboog, 2020). The core hypothesis, therefore, is that voucher-mediated contributions from the private sector can reduce inequality at scale while preserving market incentives and macro-stability—a proposition that complements, rather than replaces, traditional government programmes. In this sense, CBE should not be interpreted as an alternative welfare-state architecture, but as a complementary mechanism that operates alongside taxation, transfers, and universal public services.
A parallel trend that exacerbates the challenge of inequality is the steady increase in firm-level market power. Robust evidence indicates that average markups in both OECD and emerging economies have increased significantly above competitive benchmarks since the 1980s, shifting national income away from wages and toward profits (De Loecker & Eeckhout, 2018; Barkai, 2020; Gutiérrez & Philippon, 2017).
Rather than treating markup surplus as an unavoidable by-product of modern capitalism, Counterbalance Economics (CBE) uses the portion of price above marginal cost as a mechanism to fund a parallel purchasing system. Vouchers are issued to employees who voluntarily opt into the CBE system when it offers them greater real buying power than their current wage. These vouchers function like cash and can be used to purchase goods and services at normal retail prices, effectively raising purchasing power for lower-income workers. This mechanism allows for meaningful redistribution without extracting or taxing firm profits.
This approach also resonates with recent calls to complement progressive taxation with structural rent-capture mechanisms aimed at curbing monopoly power (Saez & Zucman, 2019; Eeckhout, Fu, & Li, 2024).
The paper first examines existing theories of income distribution alongside recent evidence on corporate markups and corporate social responsibility, building the conceptual foundation for the Counterbalance Economics (CBE) framework. It then presents an empirical strategy that should be read as an exploratory test of policy analogues rather than a definitive evaluation of a fully implemented CBE system. The empirical section therefore asks a narrower question: whether observed rent-capture and in-kind transfer arrangements resemble the distributional logic of CBE closely enough to provide preliminary evidence on its plausibility. The paper also identifies important design constraints, including inflationary transmission, firm compliance, institutional capacity, data limitations, and interaction with existing tax-transfer systems. In this sense, the contribution is primarily conceptual and policy-oriented, supported by indicative cross-country evidence that requires further validation through pilots, microdata and sector-level markup measures.
Addressing income inequality has long been a concern for economists, with an extensive body of literature documenting its causes and consequences. Classical economic theories often assume that income distribution results from marginal productivity or market forces, but even early economists acknowledged issues arising from extreme inequality. The principle of diminishing marginal utility of income (pioneered by Daniel Bernoulli and later popularised by Alfred Marshall) provides a fundamental rationale for redistribution: an extra dollar yields more utility to a poor individual than to a wealthy one. Thus, transferring resources from the rich to the poor can increase total societal welfare. This utilitarian insight laid the groundwork for policies like progressive taxation, which aims to equalise marginal utilities. However, classical theorists and successors also warned of incentive effects – taxation might dampen work and investment incentives, potentially reducing overall efficiency. Okun (1975) famously described the equity–efficiency trade-off, suggesting that redistributive policies resemble carrying water in a leaky bucket: some benefits “leak out” due to administrative costs or reduced motivation of economic agents. The prevailing policy approaches in the 20th century thus sought a balance between market-driven inequality and state intervention. Welfare states in many countries have used taxes and transfers to achieve moderate redistribution, but these mechanisms have rarely been driven to reduce inequality.
In parallel to mainstream approaches, various heterodox perspectives have expanded the discourse on inequality and distribution. Post-Keynesian economists emphasise the role of income distribution in determining aggregate demand and growth regimes. According to post-Keynesian and Kaleckian models, a more equitable distribution (higher wage share relative to profits) can lift consumption and stimulate demand-driven growth in wage-led economies (Lavoie & Stockhammer, 2013; Stockhammer, 2017). Recent cross-country evidence supports this view: Storm and Naastepad (2012) show that wage-led demand dominates in most advanced economies, while Onaran and Galanis (2014) find that raising the labour share would boost global growth even when open-economy leakages are taken into account. Empirical work by Ostry, Berg, and Tsangarides (2014) for the International Monetary Fund likewise reports that lower inequality is associated with longer and more durable growth spells, challenging the notion of an equity–efficiency trade-off. These insights suggest that reducing inequality can generate macroeconomic benefits, triggering a virtuous cycle of inclusive growth. Behavioural research further indicates that economic agents—including business owners and managers—may accept or even prefer egalitarian arrangements when social norms endorse fairness (Fehr & Schmidt, 1999; Krugman, 2012). Such tendencies could be harnessed in cooperative solutions to inequality, for instance, through voluntary voucher contributions by firms.
A growing body of evidence questions the long-held view that cash transfers are always superior to in-kind benefits. Recent impact evaluations have shown that well-designed vouchers can deliver welfare gains comparable to those of cash while achieving distributional or nutritional targets that cash sometimes misses (Cunha, 2014; Gentilini, 2016). Building on this insight, our Countertrade Credits (CCs) aim to replicate a cash-like utility within a closed corporate loop, funded by markup rents rather than the public purse. The concept draws inspiration from complementary currency and corporate-barter research. Trade exchange networks allow firms to clear transactions in non-monetary “trade credits,” mobilising idle capacity and expanding liquidity alongside official money (Johnson, 2018; Teasdale & Aiken, 2019). CBE adapts this corporate-barter logic for social purposes, redirecting a portion of aggregate markups into Community Connections (CCs) that specifically target low-income households through a nationally coordinated platform. In so doing, it marries the efficiency of private barter systems with the equity objectives of voucher programmes, forging a scalable mechanism that can ease fiscal pressure while boosting the real purchasing power of those at the bottom of the distribution.
To improve the implementation of Counterbalance Economics (CBE) by governments, it is essential to examine instances where voucher systems or similar mechanisms have been effectively used to reduce inequality without relying on traditional cash transfer methods. One notable example is the food voucher system in the United States, which has been effective in providing short-term relief to low-income families; however, it has faced criticism regarding its flexibility and efficiency compared to direct cash transfer systems (Kaldor, 1956). In many cases, these systems have been more politically acceptable because they do not rely on high taxation or direct government intervention in personal income.
It is important to emphasise at the outset that Bolsa Família and Counterbalance Economics differ fundamentally in their institutional architecture, financing, and normative foundations. Bolsa Família is a state-led, tax-financed programme embedded in a universal public policy framework with democratic accountability, whereas CBE operates through market-mediated mechanisms relying on firm-level markup rents. Another instructive example comes from Brazil’s Bolsa Família programme, the world’s largest conditional transfer scheme. Although formally classified as a cash transfer, its benefits are disbursed via an electronic card restricted to purchases of food, school supplies and basic services, creating a quasi-voucher that channels resources toward essentials while preserving household choice (Fiszbein & Schady, 2009; Soares, Ribas, & Osório, 2010). Rigorous evaluations show that Bolsa Família increases school attendance and nutrition with minimal labour-supply disincentives and at a fiscal cost of less than 0.5 per cent of GDP, making it far more resilient to inflationary pressures and leakage than untargeted subsidies (de Brauw & Hoddinott, 2011; Lichand & Oliveira, 2017). Despite these surface similarities in the use of restricted purchasing mechanisms, the institutional logic of Bolsa Família differs fundamentally from that of CBE: redistribution is affected through in-kind-oriented purchasing power rather than open-ended cash, thereby limiting macro-instability and opportunities for corruption—a point emphasised by Piketty (2014) and reinforced in Stiglitz’s (2012) call for innovative, non-distortionary transfer mechanisms. CBE extends this logic by shifting the financing burden from the public budget to the private sector’s markup rents, allowing firms to participate as complementary agents of redistribution under public oversight. At the same time, governments retain only a light supervisory role. In doing so, CBE aims to match the poverty-reduction record of programmes like Bolsa Família but with even lower fiscal strain and more substantial incentives for corporate engagement in equality-enhancing initiatives.
Thus, countries like Brazil and the United States show that voucher-based programs or similar systems are effective in improving equity. However, private models like CBE could offer a more stable and cost-effective long-term solution by aligning market incentives with social justice objectives. This could be particularly relevant for developing countries, where government capacity to finance social welfare systems is often limited (Milanovic, 2016). At the same time, a large body of empirical evidence demonstrates that universal public services—particularly in health, education, and care—are among the most effective instruments for reducing income inequality and promoting intergenerational mobility (Verbist et al., 2012; OECD, 2015). Such services not only compress income dispersion directly, but also support aggregate demand, productivity growth, and long-run human capital formation. Accordingly, any market-based or voucher-oriented mechanism should be evaluated as complementary to, rather than a substitute for, universal public provision.
Firms are now scrutinised not only for profitability but also for their environmental, social, and governance (ESG) footprints. Large-scale evidence shows that companies with high ESG scores display smaller internal pay gaps and flatter wage structures, signalling a commitment to fair labour practices (Kölbel, Heeb, Paetzold, & Busch, 2020; Hwang, Park, & Young, 2022). Investors are increasingly pricing these attributes: funds with explicit social mandates channel capital toward firms that demonstrate pay equity and inclusive-growth policies, thereby lowering their cost of capital (Eccles, Ioannou, & Serafeim, 2014; Krüger, 2015). This dynamic creates a natural opening for Counterbalance Economics (CBE). Suppose socially conscious shareholders and consumers reward companies that reduce inequality. In that case, firms gain a tangible incentive to participate in an inter-corporate voucher scheme that amplifies the purchasing power of low-income individuals.
The literature also warns, however, that voluntary corporate initiatives rarely solve systemic distributional problems unless private benefits align tightly with social goals or are supported by light regulation (Bénabou & Tirole, 2010; Liang & Renneboog, 2020). Counterbalance Economics (CBE), therefore, embeds a dual mechanism: reputational and investor rewards on the one hand, and a modest, rule-based obligation for firms to accept CBE vouchers as a parallel payment mechanism on the other. By linking participation to both market incentives and a clear regulatory framework, the model aims for a “win-win” equilibrium — firms enhance their ESG profile and benefit from increased demand, while society gains a scalable, non-tax-based tool for reducing inequality.
A further political-economy challenge is regulatory capture. The firms with the largest markups may also have the greatest lobbying capacity, legal resources and ability to redesign contracts, relocate activity or seek exemptions. CBE therefore requires anti-avoidance rules, transparent audit standards and independent oversight. Without those safeguards, a rule-based voucher obligation could be diluted by the same market power it is intended to counterbalance.
Recent research on market power shows that a growing share of corporate profits represents pure economic rent rather than compensation for productive effort (De Loecker & Eeckhout, 2018; Barkai, 2020). Rather than taxing or redistributing these rents directly, Counterbalance Economics (CBE) leverages their presence to justify issuing in-kind vouchers that rebalance purchasing power from the bottom up. Because the system operates only on the markup component—without extracting firm profits—it targets a distortion that classical tax theory treats as non-distortionary (Saez & Zucman, 2019), while potentially reducing inequality with fewer efficiency losses than conventional taxes or subsidies.
Moreover, empirical evidence from large conditional-transfer programmes shows that benefits can produce welfare gains similar to cash while avoiding the difficulties of means-testing and fraud (Fiszbein & Schady, 2009). Financing CBE vouchers through markups, therefore, aligns with the equity–efficiency frontier emphasised by Stiglitz (2012) and the IMF’s recent call for “rent-based” instruments to complement progressive taxation (Ostry, Loungani, & Furceri, 2014).
For firms, participation in a mark-up-funded voucher scheme can be framed as a profit-compatible ESG strategy. Shared-value theory predicts that companies investing in the purchasing power of low-income consumers will ultimately expand the market size and improve long-term profitability (Porter & Kramer, 2011). Evidence from capital market studies indicates that investors are increasingly rewarding firms with credible social impact commitments, thereby lowering their cost of capital (Krüger, 2015; Liang & Renneboog, 2020). If voucher obligations scale with documented markups, the burden is proportional to the firm’s market power, creating a level playing field and dampening incentives to lobby for loopholes (Gutiérrez & Philippon, 2017).
While employee participation in the CBE system is voluntary and self-sorting, firm participation should be mandated: all firms must accept vouchers as partial payment at face value. To align incentives, governments can reward compliance by recognizing voucher acceptance in ESG disclosure frameworks or offering corporate tax offsets — embedding a tangible private return for socially beneficial behavior.
Nonetheless, public intervention remains indispensable. Only the state can (i) define the benchmark competitive markup, (ii) audit firms’ cost structures, and (iii) operate a clearinghouse that redeems vouchers at par across sectors. Historical experience with education and food-stamp vouchers shows that government monitoring and standard-setting are critical for preventing quality dilution and ensuring universal acceptance (Muralidharan & Sundararaman, 2011; Chumacero & Paredes, 2015). Institutional economics research further demonstrates that rent-capture policies succeed only where regulatory capacity and legal enforcement are robust (Acemoglu & Robinson, 2012; Rodrik, 2004). Accordingly, the CBE framework envisions a public-private partnership: the Government imposes and administers the markup taxes, while firms deliver authentic goods and services through the voucher network. When these complementary roles are aligned, mark-up-financed vouchers can match—or even outperform—traditional fiscal tools in the joint pursuit of equity and growth.
This institutional requirement is particularly demanding in emerging economies. CBE would work best where governments can verify costs, monitor markups, operate a reliable clearinghouse, protect consumers from price discrimination and include informal workers through credible digital or administrative channels. Where informality is high, CBE should be piloted gradually and combined with formalisation incentives rather than assumed to scale automatically.
In the CBE framework, we propose an in-kind mechanism that utilises a portion of the aggregate markup as a signal and source for redistribution. Antitrust guidelines already treat sustained markups as potential indicators of market power (OECD, 2021). By requiring firms to accept vouchers equivalent to a share of the price above marginal cost—the “pure rent” in Auerbach’s (1979) terms—CBE targets economic surplus that can be redirected without distorting productive incentives. This approach aligns with principles from optimal rent taxation (Saez & Zucman, 2019) and mirrors the rationale behind windfall-profit mechanisms designed by the IMF (2023), though it operates through market exchange rather than fiscal extraction.
In summary, prior work underscores the importance of reducing inequality and offers insights from multiple perspectives (utilitarian welfare, Keynesian demand, behavioural fairness, and CSR/ESG motives). Nevertheless, a gap remains in how to operationalise a large-scale redistribution without relying solely on government taxation and spending. Market-driven or hybrid mechanisms have not been extensively explored in academic literature. This paper contributes to filling that gap by proposing and analysing the Counterbalance Economics framework. Building on the above strands, CBE is positioned at the intersection of economic policy and corporate strategy, introducing a new approach that leverages private sector resources (through vouchers) to achieve what traditional policies have struggled to accomplish: a substantial, sustainable reduction in income inequality.
This study uses an exploratory quasi-experimental panel design to examine whether policy arrangements that partially resemble CBE are associated with weaker links between markups and income inequality. We rely on a two-way fixed-effects difference-in-differences framework, while explicitly recognising the limits of causal identification in this setting. The 1982-2016 sample contains only three within-sample treated cases that can contribute to identification: China, Malaysia and Australia. Germany, France, Sweden and Brazil adopted relevant windfall or rent-capture measures after 2016 and are therefore retained only as external policy analogues in the discussion and coding rationale, not as treatment observations in the estimating sample. As Angrist & Pischke (2009) note, difference-in-differences can be understood as a form of fixed-effects estimation, but its credibility depends on the parallel-trends assumption and sufficient treatment variation. We therefore interpret the estimates as indicative evidence on rent-capture mechanisms rather than a conclusive causal estimate of CBE itself.
Our identification strategy is grounded in the potential-outcomes framework (Rubin, 2005; Imbens & Rubin, 2015), which defines causal effects as contrasts between outcomes under alternative treatment states. We denote Y_it(0) as country i’s inequality at time t in the absence of a CBE-like policy and Y_it(1) as the corresponding outcome under a CBE-like policy analogue. The treatment effect is therefore Y_it(1) - Y_it(0), although only one potential outcome is observed for each country-year. In this paper, the treatment indicator does not measure full CBE implementation. It captures narrower policy analogues in which governments recoup part of sectoral economic rents and channel them toward in-kind transfers, subsidies or consumer-price offsets. This distinction is central: the estimates speak to the empirical plausibility of rent-capture redistribution, not to the full inter-corporate voucher system proposed in the theoretical framework.
This context also involves measuring how markups (a proxy for firms’ market power) affect inequality. We exploit panel causal inference methods to disentangle correlation from causation: the inclusion of fixed effects helps control for unobserved country characteristics, and we assume policy timing is exogenous or unrelated to short-term inequality shocks (justified later). Additionally, by interacting with the CBE policy with markups, we identify how the policy’s introduction alters the link between corporate pricing power and inequality. We ground our empirical methodology in established econometric literature, drawing on insights from Angrist & Pischke (2009) on causal research design, Imbens & Rubin’s (2015) formalisation of treatment effects, and Wooldridge’s (2010) panel data techniques for robust inference. In summary, the methodological approach blends difference-in-differences logic with panel econometric tools to estimate causal effects, guided by best practices from the so-called “credibility revolution” in empirical economics (e.g., transparency in assumptions and robust checks).
To construct the CBE analogue indicator (CBE_it), we reviewed policy actions across the 12 countries to identify cases where (a) firms or sectors were required or incentivised to share economic rents, and (b) the resulting value was channelled through in-kind transfers, targeted consumer subsidies, vouchers or price offsets. We use the term analogue deliberately. No country in the sample has implemented CBE in its full theoretical form, and the observed policies are closer to government-administered windfall or rent-capture instruments than to a firm-level markup-funded voucher system. The indicator should therefore be interpreted as a proxy for the rent-recycling logic of CBE, not as a direct measure of CBE adoption.
Based on this classification, seven countries are relevant for the conceptual coding discussion, but only three provide within-sample treatment variation for the 1982-2016 estimates:
• Germany and France introduced EU-aligned windfall levies on energy firms in 2022, using the proceeds to fund energy price caps and direct support to households (Reuters, 2022; OECD, 2023).
• Sweden implemented the EU-mandated solidarity contribution in 2023, committing the revenues to targeted financial relief for vulnerable consumers (Swedish Government, 2023).
• Australia’s Minerals Resource Rent Tax (MRRT), though repealed, explicitly financed family and education-related transfers, including school vouchers (Richardson, 2012; Treasury of Australia, 2013).
• In 2023, Brazil imposed a temporary export tax on crude oil to fund domestic fuel subsidies, thereby directly transferring windfall profits to protect consumers (Folha de São Paulo, 2023).
• China has had a price-triggered windfall tax on oil producers since 2006, with revenues funding national fuel subsidies (IMF, 2011; Xinhua, 2006).
• Malaysia imposes a Windfall Profit taxes on palm oil companies, with revenues used to stabilise cooking oil prices for consumers (Malaysian Ministry of Finance, 2022).
In contrast, Japan, Canada, Mexico, South Africa, South Korea and Ireland either did not adopt a comparable windfall-rent instrument within the sample window or did not earmark revenues for voucher-style or consumer-price redistribution. This distinction partially captures the fiscal and distributive logic of CBE, but it does not eliminate construct-validity concerns: most observed cases concern sector-specific resource or energy rents rather than economy-wide firm-level markups. The empirical analysis should therefore be read as a test of CBE-adjacent rent-recycling policies rather than a direct test of the proposed CBE voucher architecture (see Table 1).
Our analysis uses an unbalanced panel dataset of 12 OECD and emerging economies observed annually from 1982-2016. The countries were selected based on data availability and variation in development levels, ensuring representation of advanced and middle-income economies. The dataset is appropriate for exploratory analysis, but it is not without limitations: the identifying variation is concentrated in three within-sample treated cases, and the Gini series is incomplete for several country-years. These features reduce statistical power and require caution in interpreting the DiD estimates. The data are compiled from internationally recognised sources, and key variables are constructed as follows:
• Income Inequality (Gini Index): Measured as the Gini coefficient of disposable household income (post-tax, post-transfer) for each country year. We obtain Gini estimates from World Bank and LIS-harmonised sources where available. The Gini index is expressed on a 0-100 scale (0 = perfect equality). The assembled panel contains 200 observed Gini values out of 420 possible country-years, implying substantial missingness. Because interpolation can mechanically smooth pre-treatment trends and may be especially consequential in a DiD design, interpolated values are used only where short gaps can be justified, and the sensitivity of the results to non-interpolated and balanced-window specifications is treated as a core robustness requirement rather than a minor data-cleaning step.
• Markup (aggregated): The economy-wide average corporate Markup, defined as the ratio of price to marginal cost. We utilise the dataset by De Loecker and Eeckhout (2018) and updates thereof, which provide estimates of markups for many countries. These estimates capture broad trends in market power; for example, prior research finds that markups have risen 30–60% above competitive levels in OECD economies since the 1980s. We express the aggregate markup level as a number (e.g., 1.00 = perfect competition; 1.60 = 60% markup over cost). To focus on market power, we derive a variable below.
• Counterbalance Economics Policy (CBE_it): This binary indicator reflects the presence of policy actions analogous to the CBE framework. For each country i, CBE_it = 1 from the first year that the country implements a policy requiring firms to share economic rents—such as markups or windfall profits—via in-kind mechanisms earmarked for redistribution (e.g., vouchers, price subsidies, or consumer credits), and remains 1 thereafter. While these cases do not constitute full implementation of CBE, they capture core structural elements of the model.
• Specifically, CBE_it = 1 in years when a country has an active policy that mirrors key features of the CBE model: namely, requiring or incentivizing firms to share excess profits or economic rents (such as markups or windfalls) and redistributing value through in-kind mechanisms targeted at lower-income households—such as vouchers, consumer credits, or utility price offsets. In our sample, a subset of countries adopted such analogous policies, though none implemented full-scale CBE. For example, if Italy introduced a profit-sharing or in-kind redistribution policy in 2010, then CBE_it = 0 for Italy up to 2009, and CBE_it = 1 from 2010 onward. Countries without such policies by 2016 remain at zero throughout. Cases like Italy’s 2017 profit-in-kind pilot (outside our sample window) and similar initiatives elsewhere inform our coding approach.
We interpret these as proxies for CBE-adjacent policy design. However, the staggered adoption structure is weaker than in a standard multi-period DiD design because only China, Malaysia and Australia enter treatment before 2016. Post-2016 cases are not used for identification. Data on policy adoption years are sourced from government reports and policy databases; when exact enactment dates are available, the indicator is coded from the first full year of implementation. Because CBE is a novel concept, the few treated cases should be understood as partial analogues involving rent capture and in-kind or price-offset redistribution, not as implemented examples of the full CBE mechanism.
• Macroeconomic Control Variables: To isolate the impacts of markups and CBE on inequality, we include a rich set of controls:
○ GDP per capita (PPP, constant 2021 international $): Logged GDP per capita to capture the level of economic development and income from the World Bank World Development Indicators. Higher GDP per capita could affect inequality through various channels (Kuznets effects, etc.), so it is important to control for overall prosperity.
○ Net Barter Terms of Trade Index (2015 = 100): The terms of trade, as reflected in World Bank data, represent external economic conditions (export prices vs. import prices). Fluctuations in commodity prices or trade conditions can influence income distribution (e.g. through wage and profit shifts in tradable sectors).
○ Research & Development Expenditure (% of GDP): R&D intensity, from the World Bank or OECD, as an indicator of technological progress and innovation-driven growth. This may proxy for the knowledge economy’s rise, potentially affecting wage dispersion (skilled vs unskilled labour).
○ Unemployment Rate (% of the labour force): From the International Labor Organization or WDI to control for labour market slack. High unemployment can suppress wages for lower-skilled workers and affect inequality. It also captures business cycle effects beyond GDP.
○ Trade Openness (% of GDP): Although not explicitly listed in the prompt, the trade/GDP ratio is often included (some data descriptions mention trade openness and tax revenue). If relevant, we include openness to account for globalisation’s effects on inequality.
Each control variable is drawn from reputable databases, but cross-country comparability and missingness remain important limitations. Some macroeconomic controls have gaps in early years and for emerging economies, while the Gini variable is observed in fewer than half of all possible country-years. Remaining missing values are handled through listwise deletion in the main specifications, with interpolation used only in sensitivity checks where short gaps are clearly bounded. The summary statistics in Table 2 should therefore be read as descriptive diagnostics for the assembled panel, not as evidence that the panel is balanced. The zero-coded GDP per capita observation identified in the raw extraction is treated as a missing-data placeholder and excluded prior to log transformation; no logarithm is taken over zero values. This correction is essential because otherwise the log-GDP regressor would be mechanically invalid.
We take the natural log of GDP per capita because it linearises the otherwise nonlinear link between economic development and inequality, making the regression fit more reliable. Expressing income in logs converts the coefficient into an elasticity, so a 1 per cent rise in real income per person has a directly interpretable percentage effect on the Gini. The log transformation also stabilises the variance—dampening the influence of very rich or impoverished countries—and thus reduces heteroskedasticity in the error terms.
We control for trade openness rather than poverty because external competition directly disciplines domestic markups—the financing base of CBE—whereas poverty is essentially an outcome variable.
Countries with within-sample CBE-adjacent adoption are treated as the treatment group, while the rest serve as controls. The key identification assumption is that, in the absence of these policies, treated and untreated countries would have followed parallel trends in inequality. Because the number of treated countries is small and missing Gini data are non-trivial, the parallel-trends assessment is reported as a diagnostic rather than as definitive proof of causal validity.
Trade openness is less collinear with the Gini coefficient than head-count poverty, thereby reducing the risks of endogeneity and multicollinearity in the regression.
Finally, complete and comparable trade-to-GDP series are available for all OECD countries in our panel, while poverty data are patchy and would shrink the sample and statistical power.
The econometric model builds on the theoretical framework linking markups to income inequality. Higher markups, defined as the wedge between prices and marginal costs, may increase inequality by shifting surplus toward firms, owners and high-income households while weakening wage growth and consumer purchasing power. The CBE framework proposes that part of this surplus could finance voucher-based redistribution. Empirically, however, the available policy analogues are closer to government-administered rent-capture instruments than to firm-level voucher obligations; the model should therefore be interpreted as a reduced-form test of rent-recycling logic rather than a structural test of the full CBE design.
The baseline econometric model is as follows:
The coefficients of interest are:
• \beta_1 on Markup_{it}: This measures the relationship between markups and the Gini index in the absence of the CBE policy. We expect \beta_1 > 0, indicating that higher markups (greater market power rents) are associated with higher income inequality. The intuition is that when firms extract monopoly rents, the surplus primarily accrues to owners or top executives, while consumers (and possibly workers) get a smaller share, thus widening inequality. Our model treats this as a baseline effect.
• \beta_2 on CBE_{it}: This captures the direct effect of implementing CBE on inequality when Markup = 0. If a country with no markups (hypothetically) enacts CBE, \beta_2 would be the expected change in Gini. In practice, if Markup is minimal, this term reflects any level shift in inequality due to the policy (for example, introducing vouchers might slightly reduce inequality, even with tiny rents, via redistribution). We anticipate \beta_2 \le 0 (CBE should not increase inequality; it might have a slight adverse effect by transferring some income to people with low incomes). However, this term may be statistically insignificant if the policy only operates meaningfully when there are rents to redistribute.
• \beta_3 on the interaction Markup \times CBE: This is the key coefficient for our hypothesis. It measures how the effect of markups on inequality is altered in the presence of a capital-based economy (CBE). A negative \beta_3 would indicate that when CBE is active, the inequality-increasing impact of markups is reduced (or even reversed) because the policy captures part of those rents for redistribution. In other words, \beta_3 quantifies the extent to which markup rents recycled through vouchers lower inequality. We expect \beta_3 < 0 and sizable in magnitude (for instance, our preliminary findings suggest that each 0.1 point of markup rent taxed and redistributed lowers the Gini by about 0.01, a meaningful effect). The combination of \beta_1 and \beta_3 allows us to compute the net effect of markups on inequality under CBE: \beta_1 + \beta_3 would be the marginal effect of Markup when CBE=1. Ideally, \beta_3 in absolute value should be close to \beta_1, which would imply CBE can neutralise the inequality impact of markup rents. A more negative \beta_3 (greater in magnitude than \beta_1) would suggest CBE overcompensates, possibly even reducing inequality on the net as markups rise.
• \mathbf{\Gamma}’\mathbf {X}_{it}: This represents the vector of coefficients on control variables \mathbf {X}_{it}. These controls (GDP per capita, terms of trade, R&D, unemployment, etc.) are included to account for other factors that influence inequality. For instance, higher GDP per capita might correlate with both higher markups and evolving inequality (the Kuznets curve), so controlling for GDP ensures \beta_1 is not confounded by development level. We expect certain controls to have significant effects (e.g., unemployment may increase inequality, while GDP growth might reduce it modestly, etc.); however, our primary interest lies in isolating the coefficients on markups and CBE.
The functional form is linear in levels, which is standard for panel regression analyses of the Gini index. Although the Gini coefficient is bounded between 0 and 100, in practice, the values typically lie in a mid-range where the linear approximation is reasonable; we verified that the predicted Gini values remain within a feasible range. We choose a linear model to ease the interpretation of coefficients as approximate percentage-point changes in inequality. The additive fixed effects soak up unobserved heterogeneity under the assumption that such heterogeneity is time-invariant. We considered whether a random-effects model might be appropriate. However, a Hausman specification test (Hausman, 1978) decisively rejected the random effects assumption (p < 0.01), indicating that country effects are correlated with our regressors. This justifies the use of fixed effects, as it provides consistent estimates when regressors (like GDP or markups) are endogenous to latent country traits.
It is essential to note that equation (1) is a two-way fixed-effects difference-in-differences estimator applied to a staggered and sparse adoption setting. Recent advances in econometric theory (e.g., Callaway & Sant’Anna, 2021) have highlighted that standard two-way fixed-effects DiD can be biased if treatment effects vary over time or across groups. Because treatment in the estimating window is limited to China, Malaysia and Australia, we report the baseline model as the main exploratory specification and treat dynamic/event-study estimates as diagnostics with limited power. The model assumes a common-trends counterfactual: absent CBE-adjacent policies, inequality in treated countries would have followed a similar path to that in control countries, conditional on controls.
Additionally, to ensure our functional form is not misspecified, we will test alternative specifications, e.g. allowing nonlinearity in Markup (quadratic term) or using the level of markups (not just) interacted with CBE. We will also explore dynamic effects by estimating leads and lags of the CBE treatment (using an event-study analysis) to verify that there are no pre-treatment effects (placebo leads should be zero) and to observe the evolution of the treatment effect over time. If pre-trend violations are detected, we may adjust the model (e.g., by including country-specific trends or using matching on trajectories) to improve identification.
The theoretical rationale behind this model is grounded in both classical and contemporary economic theory. Classical incidence analysis suggests that a tax on profits if redistributed, should reduce inequality without distorting marginal production decisions (since it targets rents). We embed that idea by focusing on Markup. Heterodox theories (e.g., post-Keynesian) posit that functional income distribution (wages vs profits) drives personal inequality; thus, rising markups (profit share) increase inequality. CBE seeks to counteract this by reallocating part of profits to lower-income households, which our interaction term captures. In essence, equation (Acemoglu & Robinson, 2012) can be viewed as a reduced-form representation of a structural model, where inequality equals f (market power, redistribution policy, other factors). We avoid a simultaneous equations approach by treating policy adoption as exogenous or at least predetermined relative to short-run inequality shocks (justified, for example, if political decisions to implement CBE respond to long-run inequality trends or external pressures rather than year-to-year noise). Still, we remain cautious about endogeneity: if high inequality itself triggers CBE adoption, then CBE_it might be endogenous. We will address this concern using checks such as instrumental variables (if a valid instrument for CBE timing can be identified, perhaps through political shifts) or matching methods to ensure that treated and control countries were on similar inequality trajectories before adoption. At a minimum, the event-study plots and parallel trend tests will be reported to support the causal interpretation of \beta_2,\beta_3.
Equation (1) is estimated by ordinary least squares with country and year fixed effects. Under the stated assumptions, this provides an estimate of the association between markups, CBE-adjacent treatment and inequality. We avoid describing the coefficient as an unbiased causal effect because the treatment proxy is imperfect, the number of treated countries is small, and missing outcome data may affect pre-trend diagnostics. Robust and clustered standard errors, alternative missing-data treatments, non-interpolated specifications, and sensitivity checks with lagged inequality are therefore reported as robustness exercises rather than as complete solutions to identification concerns.
To strengthen the empirical assessment, we conduct placebo tests by randomly assigning adoption dates, compare estimates using interpolated and non-interpolated Gini series, and examine pre-treatment trends where sufficient observations exist. Dynamic DiD and synthetic-control checks are reported as complementary diagnostics, particularly for early adopters such as Malaysia (2008), but the small number of treated units limits their inferential strength. These checks reduce the risk of spurious correlations but do not fully eliminate concerns about construct validity, sparse treatment variation or missing outcome data.
Thus, the analysis employs ordinary least squares (OLS) with two-way fixed effects (country and year), utilising clustered standard errors at the country level to address autocorrelation and heteroscedasticity across temporal observations within countries (Petersen, 2009). Rigorous diagnostic tests validate the model’s robustness: the Breusch-Pagan test (Breusch & Pagan, 1979) detects heteroskedasticity, the Wooldridge test (2010) identifies autocorrelation in panel data, and the Variance Inflation Factor (VIF) analysis screens for multicollinearity among explanatory variables.
The results suggest that higher markups are associated with greater inequality and that this association is weaker, and in some specifications reversed, in country-years with CBE-adjacent rent-recycling policies ( Table 3).
The analysis reveals three main patterns. First, markups exhibit a positive relationship with inequality: a 1-unit increase in markups is associated with a 12.34-point increase in the Gini coefficient (*p* = 0.002). Second, the interaction term Markup x CBE is negative and statistically significant (-15.22, *p* = 0.010), suggesting that rent-recycling policy analogues may weaken the inequality-enhancing association of markups. Third, control variables show mixed effects: higher GDP per capita is associated with lower inequality (-1.02, *p* = 0.034), while unemployment is associated with higher inequality (+0.42, *p* = 0.003). Terms of trade and R&D intensity are not statistically significant in this specification.
Robustness checks broadly preserve the sign of the interaction, but the results should not be read as definitive proof that CBE itself reverses inequality. The standalone CBE coefficient is insignificant (-0.87, *p* = 0.195), consistent with the theoretical claim that the mechanism operates through rent recycling rather than a simple policy-level shift. The net coefficient (*12.34-15.22 = -2.88*) should be interpreted as the estimated marginal association under the analogue policy proxy, conditional on the model and sample limitations.
The subgroup estimates in Table 4 suggest larger interaction coefficients for emerging economies than for developed economies. However, these estimates are especially sensitive to treatment definition, missing observations and institutional heterogeneity. Rather than demonstrating that CBE is automatically more effective in emerging economies, the results indicate that countries with higher initial inequality may have greater potential gains from rent-recycling mechanisms if the necessary reporting, enforcement and voucher-acceptance infrastructure exists.
The effectiveness of CBE critically depends on institutional capacity, regulatory enforcement, and sustained firm participation; absent these conditions, its distributive impact is likely to be limited. Our analysis assumes firms honour vouchers at existing prices, idle capacity prevents inflationary pressure, and voucher receipts translate into real consumption. These assumptions are restrictive. A large increase in purchasing power for low-income households, whose marginal propensity to consume is high, could generate demand-side pressure in non-tradeable sectors such as housing, care, local transport and personal services. CBE would therefore require explicit inflation safeguards, phased implementation, supply-side monitoring and rules that prevent firms from raising posted prices while accepting vouchers. The simulation also treats agents as homogeneous, ignoring stigma, unequal access to voucher-accepting outlets, and heterogeneity in local market capacity. Measuring in-kind benefits will likely require new surveys, multiple inequality indicators and revised reporting systems. Because CBE is untested at scale, small pilots would be necessary to observe how firms negotiate voucher supply, how employees value vouchers relative to cash, and whether the mechanism affects prices, employment or investment.
For the econometric approach, robustness can be assessed by using alternative inequality measures, such as the Theil index or Atkinson index, as outcomes to confirm that results are consistent across different definitions of inequality. We could also envisage a placebo test: use a falsified treatment date (when no CBE had been implemented yet) to verify that no “effect” is found before the real intervention. Additionally, a synthetic control method could complement DiD for any single treated unit of interest (e.g., if one country or state adopts CBE at a time, creating a synthetic version of that unit from the data of others to serve as a comparison). Another empirical improvement would be to incorporate survey data on households, capturing how consumption patterns change with voucher receipts – for example, do recipients treat vouchers exactly like cash, or do they spend differently (perhaps more on essentials)? Such microdata analysis would enrich the understanding of CBE’s welfare implications beyond what macro indicators show. Lastly, future research might integrate a macroeconomic model (such as a Computable General Equilibrium model) to simulate long-run effects on investment, productivity, and government budgets, especially if CBE allows reallocation of public spending (for instance, reducing welfare expenditure if the private system covers basic needs). These extensions are beyond the scope of the current paper but are important for a complete evaluation of the policy’s viability.
A further limitation concerns long-run growth accounting. Some markups may reflect temporary rents, while others may partly finance fixed costs, R&D recovery and innovation. A poorly designed rent-recycling obligation could reduce investment incentives if it fails to distinguish monopoly rents from innovation-related quasi-rents. Future work should therefore model dynamic effects on investment, productivity, firm entry and the cost of capital, rather than focusing only on short-run distributional outcomes.
CBE must also be evaluated in relation to existing tax-transfer systems. The mechanism could complement progressive taxation, VAT exemptions, social insurance and means-tested transfers by adding purchasing power without replacing universal services. Conversely, it could crowd out political support for broader public provision if presented as a substitute for the welfare state. Policy design should therefore specify whether CBE is additive, budget-neutral or partially substitutive, and should include safeguards against reducing existing entitlements.
Finally, relying on aggregate markups masks whether rents stem from a few concentrated sectors or are widely dispersed. Ignoring industry heterogeneity may understate distributional gains where excess profits cluster in limited sectors and obscure sector-specific responses, including innovation, lobbying, regulatory arbitrage, exit, and price changes. This issue is particularly important because high-markup firms may also possess stronger political influence and greater capacity to resist, dilute or reshape regulation. Future studies should merge firm- or industry-level markup data with micro evidence on voucher contributions and take-up to evaluate heterogeneous effects, compliance incentives and the design of sector-targeted obligations.
Our empirical analysis provides indicative evidence that corporate markups are positively associated with income inequality across OECD and emerging economies, consistent with theoretical predictions from recent economic literature (De Loecker & Eeckhout, 2018; Autor et al., 2020). Specifically, the preferred specification estimates that a 10-percentage-point increase in markups above a competitive benchmark is associated with an increase of approximately 1.23 points in the Gini index. This magnitude is economically meaningful, but the result should be interpreted cautiously because outcome missingness, sparse treatment variation and the analogue nature of the policy proxy limit causal claims. The finding nevertheless aligns with studies suggesting that market concentration and corporate pricing power can exacerbate economic disparities (Philippon, 2019; Barkai, 2020).
Crucially, our findings indicate that, under specific institutional and regulatory conditions, the Counterbalance Economics (CBE) intervention can neutralise—and in some contexts partially reverse—the inequality-widening effects of corporate market power.
The interaction term between Markup and the CBE-adjacent policy indicator is negative and statistically significant in the preferred specification and remains directionally consistent across several robustness checks. Quantitatively, when the policy proxy is active, the marginal association between additional markups and inequality is estimated to diminish and, in the baseline model, turn negative (net slope approximately -0.29 per 0.10 markup). This empirical pattern is consistent with the central proposition that recycling economic rents toward lower-income households can mitigate markup-driven inequality, but it should not be read as definitive validation of a firm-funded voucher system because the observed policies are only partial analogues of CBE.
The robustness analyses provide useful but not conclusive support for the interpretation of the results. Placebo adoption dates generate smaller and statistically weaker interaction effects, while event-study diagnostics do not reveal clear pre-treatment divergences where data are available. However, because Gini observations are missing for a substantial share of the panel and the number of treated countries within the sample is limited, these exercises cannot fully rule out residual confounding. The results therefore support a cautious policy interpretation: strategically designed rent-capture mechanisms, coupled with targeted in-kind redistribution, may help offset inequality associated with corporate market power, but direct evidence on CBE requires pilots, microdata and fully reproducible empirical materials.
Redistributive policies such as CBE have gained attention because inequality remains high and traditional tax-transfer systems face political, fiscal and administrative constraints. Yet the same institutional weaknesses that motivate CBE in emerging economies can also limit its feasibility. In economies with large informal sectors, many low-income workers and small firms may operate outside formal payroll systems, markup reporting, digital clearinghouses or voucher-acceptance networks. CBE would therefore require complementary institutional investments: formalisation incentives, verifiable cost and markup reporting, anti-avoidance rules, consumer-protection standards, and integration with existing fiscal and social-protection systems. ESG incentives may support compliance, but they are unlikely to be sufficient without enforceable rules and credible monitoring.
These results highlight the potential of CBE as a policy lever for redistribution, but they also require a more cautious interpretation than a direct policy-effect claim. By systematically recycling a portion of economic rents through vouchers or targeted in-kind transfers, CBE could offset some inequality-enhancing effects of corporate market power. The negative interaction between Markup and the policy analogue suggests that such benefits may scale with markup levels, making the mechanism especially relevant in concentrated markets. However, the estimates are based on partial policy analogues, sparse within-sample treatment and incomplete outcome data. Non-significant variables such as R&D may reflect measurement limitations, context-specific dynamics or offsetting channels, including the possibility that markups finance fixed-cost recovery and innovation.
The subgroup results suggest that CBE-adjacent mechanisms may have greater distributional potential in economies with higher initial inequality, but this does not imply automatic feasibility in emerging economies. The same contexts often have larger informal sectors, weaker administrative capacity and lower compliance with formal reporting systems. CBE should therefore be viewed as potentially valuable only where governments can monitor markups, enforce voucher acceptance, prevent price discrimination and coordinate the scheme with existing tax, VAT, social-insurance and means-tested transfer systems.
CBE utilises transfer vouchers to provide additional purchasing power for essential goods and services, but it cannot be assumed to be neutral with respect to prices, innovation incentives or existing tax-transfer systems. Its promise lies in weaving redistribution into market transactions while maintaining public oversight and preserving universal public provision. The revised analysis therefore positions CBE not as a proven substitute for fiscal redistribution, but as a framework worthy of cautious empirical testing. By mobilising private-sector resources in a rule-based and supervised architecture, CBE may contribute to mitigating markup-driven inequality where traditional redistributive capacity is constrained. Its role, however, must remain complementary to progressive taxation, VAT design, social insurance, means-tested transfers and universal services. Further research should prioritise pilots, sector-level markup evidence, long-run growth accounting and open replication files before strong policy conclusions are drawn.
| Views | Downloads | |
|---|---|---|
| F1000Research | - | - |
|
PubMed Central
Data from PMC are received and updated monthly.
|
- | - |
Provide sufficient details of any financial or non-financial competing interests to enable users to assess whether your comments might lead a reasonable person to question your impartiality. Consider the following examples, but note that this is not an exhaustive list:
Sign up for content alerts and receive a weekly or monthly email with all newly published articles
Already registered? Sign in
The email address should be the one you originally registered with F1000.
You registered with F1000 via Google, so we cannot reset your password.
To sign in, please click here.
If you still need help with your Google account password, please click here.
You registered with F1000 via Facebook, so we cannot reset your password.
To sign in, please click here.
If you still need help with your Facebook account password, please click here.
If your email address is registered with us, we will email you instructions to reset your password.
If you think you should have received this email but it has not arrived, please check your spam filters and/or contact for further assistance.
Comments on this article Comments (0)