Tax progressivity and taxing the rich in developing countries: lessons from Latin America

Abstract

This article discusses the challenges and potential policy choices for levying progressive taxes and taxing the rich in Latin America, a region known for its high-income inequality, limited tax-collection capacity, and low share of taxes collected from personal income and wealth. Factors such as high exemption thresholds, low top marginal tax rates, and limited administrative capacity undermine the redistributive ability and revenue collection of the tax systems in the region. Moreover, the income composition for the top percentiles largely comes from capital, and the effective tax rates they face are often low due to the preferential treatment of capital income and wealth. After discussing the evidence of how the rich in Latin America respond to progressive taxes on income and wealth and changes in enforcement policy, we provide some insights on potential policy choices to tax them effectively. These may include broadening the income tax base by lowering the number of exempt and non-taxable income items and the statutory exemption thresholds, reevaluating preferential tax rates on capital income, monitoring foreign income, addressing the abuse of tax treatment by business earners, and enhancing tax administration capacity. Additionally, wealth taxes may complement the tax system with updates to property registers and scrutiny of foreign assets.

I. Introduction

This paper describes developing countries’ experiences of levying progressive taxes and taxing the rich. Progressive personal income and wealth taxes could help these countries address their high inequality and raise tax revenue. However, governments in these countries face specific challenges in terms of administration and enforcement due to factors such as informality, limited administrative resources, and weak enforcement capacity. We approach this study by focusing on the experience of Latin American (LA) countries. We focus on this region for two main reasons. First, it is a region with high inequality and wealthy individuals are as rich as in some European countries. Second, high-quality administrative data and policy variation have fostered new research on tax progressivity and taxing the rich.

The paper is organized as follows: section II presents stylized facts; section III discusses factors eroding revenue collection and the redistributive capacity of personal income tax (PIT) systems; section IV focuses on the income composition and tax burden of wealthy individuals in LA countries: section V discusses the challenges of taxing capital in these countries; section VI reviews evidence of how the rich respond to progressive taxation and enforcement; and section VII offers insights as to policy choices to improve tax progressivity and effectively tax the rich in developing countries.

II. Stylized facts

This section presents an overview of income and wealth concentration in Latin America, the tax structure, and the redistributive capacity of the PIT systems. We use well-known databases and standard measures to compare developing and developed regions.¹

Fact #1: Latin America is one of the most unequal regions in the world.

Since our focus is on the rich, we compare estimates of the total income and wealth shares held by the top percentiles of the income and wealth distributions. We base our analysis on LA countries where these estimates are available. While there are considerable data limitations and measurement challenges in estimating top shares, especially in developing countries, we will take these estimates at face value for now and discuss the issues in section IV.

Figure 1 presents the top 1 per cent income and wealth shares across world regions (panel (a)) and LA countries (panel (b)) for the year 2020. The figure shows that wealth is more concentrated than income and that income and wealth are highly concentrated in the developing world. For instance, the top 1 per cent income (wealth) share is 20 per cent (35 per cent) in the LA region compared to only 12 per cent (26 per cent) in Europe. However, there are substantial disparities across countries. For example, the top income share in Mexico, Chile, and Brazil roughly doubles that of low-inequality countries like Argentina and Uruguay, which stand closer to the European average.

Figure 1:

Top 1 per cent income and wealth shares, 2020. (a) World Regions. (b) LA countries.

Notes: This figure shows the share of total income and net wealth earned by the top 1 per cent of adults across world regions (panel (a)) and LA countries (panel (b)) for 2020. The black bars display income shares, while the grey bars report wealth shares. Income is defined as pre-tax national income measured after pension and unemployment contributions and benefits paid but before income taxes and other transfers. Wealth is defined as net household wealth among adults, including financial and non-financial assets owned by individuals, net of their debt. The Latin American average in panel (a) represents the unweighted average of the 10 LA countries included in panel (b).

Source: Own elaboration based on World Inequality Database (https://wid.world/). Table 1, in the Appendix, lists countries grouped by world regions. For methodological details, see https://wid.world/methodology/.

Fact #2: Latin America has a low tax-collection capacity compared to more developed countries.

Figure 2 shows the total average tax revenues as a share of the gross domestic product (GDP) across world regions (panel (a)) and LA countries (panel (b)) for the year 2020. Despite recent changes (Bachas et al., 2022), there is about a 7 percentage point gap in the tax-to-GDP ratio between Europe and Latin America.² Figure 2(b) shows that tax-to-GDP ratios vary across LA countries, from 13.2 per cent (Peru) to 23.8 per cent (Argentina), but most fall below expected ratios for middle- and upper-middle-income countries.

Figure 2:

Tax revenue and tax structure, 2020. (a) World Regions. (b) LA countries.

Notes: This figure shows the tax revenue as a share of the gross domestic product across world regions (panel (a)) and LA countries (panel (b)) for 2020, separated by revenue sources. Indirect revenue source (in light grey) is defined as total tax revenue minus social security contributions (item 2000) and direct tax revenue (items 4000 + 1000). Direct tax revenue is separated into three components: personal taxes on income and capital gains, taxes on property, wealth, and wealth transfers, and taxes on corporate income. The Latin American average in panel (a) represents the unweighted average of the 10 LA countries included in panel (b).

Source: Own elaboration based on OCDE.stats (https://stats.oecd.org/). Table 1, in the Appendix, lists countries grouped by world regions.

Fact #3: Latin America collects a smaller share of taxes from income and wealth.

Figure 2 shows the tax-to-GDP ratio by revenue source: (i) corporate income taxes, (ii) personal taxes on income and capital gains, (iii) property, wealth, and wealth transfer taxes, and (iv) ‘indirect’ taxes, mainly consumption taxes. Latin America relies heavily on consumption taxes like the value-added tax (VAT) compared to OECD countries, where taxes on income and wealth play a dominant role. In particular, the importance of taxes on personal income and capital gains is considerably smaller than in developed regions. This pattern has not changed despite an increase in tax-to-GDP ratios in recent decades (Barreix et al., 2017; Bachas et al., 2022).

Fact #4: The small size of income taxes in Latin America weakens the redistributive capacity of its tax system.

Income taxes in Latin America are generally progressive but do not meaningfully reduce income inequality due to the meagre amount of revenue they collect (Hanni et al., 2015; Barreix et al., 2017; Lustig, 2017; Clifton et al., 2020; Bachas et al., 2022).

Figure 3 compares the redistributive capacity of the personal income tax (PIT) across world regions (panel (a)) and LA countries (panel (b)). The grey bars plot the redistributive power using the Reynolds–Smolensky index, which splits the variation of the tax’s redistributive capacity into a portion caused by changes in the average effective tax rate (black hollow squares) and another part caused by changes in its progressive capacity (black crosses), measured by the Kakwani index of progressivity of the PIT schedule³ Panel (a) shows that Europe, Oceania, and North America have a higher redistributive capacity than Asia, Africa, and Latin America. The Latin American PIT schedule is progressive but has a weak redistributive capacity due to its low average tax rate. Indeed, panel (b) shows that LA countries generally display a strong progressive capacity but frail redistributive capacity due to the small size of their PIT (with some exceptions, such as Argentina). Thus, the redistributive capacity remains low in LA primarily due to its small contribution to taxes, rather than its statutory progressivity.

Figure 3:

Personal income tax: redistributive capacity as a function of progressive capacity and the aggregate average tax rate, 2020. (a) World Regions. (b) LA countries.

Notes: This figure shows the redistributive capacity of the PIT across world regions (panel (a)) and LA countries (panel (b)) for 2020. The bottom x-axis displays the Reynolds–Smolensky index of the redistributive power of the PIT system (grey bars). A higher bar means higher redistributive capacity. The upper x-axis splits the two components of the Reynolds–Smolensky index: the Kakwani 1977 index of the progressivity of the PIT schedule (black crosses) and the aggregate average tax rate (black hollow squares).

Source: Own elaboration based on tabulates from Vellutini and Benitez (2021). Table 1, in the Appendix, lists countries grouped by world regions.

In summary, Latin America is one of the most unequal regions in the world, and its tax-collection capacity and tax mix differ significantly from developed regions. The small size of progressive income taxes weakens the redistributive capacity of the tax system, but raising the average effective tax rate can help address this issue.

III. Factors undermining tax progressivity

The previous section showed that taxes on income and wealth are some of the most progressive tax instruments. Among OECD countries, the PIT is a primary source of revenue. However, in Latin America, the PIT falls significantly below OECD standards, collecting only about 2 per cent of GDP (OECD et al., 2022). This section outlines the factors that undermine tax progressivity and the redistributive capacity of the tax system. Specifically, it examines both tax and non-tax features, such as the exemption threshold, the top marginal tax rate (TMTR), tax allowances, evasion, administrative capacity, work and business informality, and tax morale.

The PIT exempts most individuals and has low TMTR

Figure 4 compares the PIT’s statutory TMTR and the exemption threshold relative to gross national income per capita (GNIpc) for the US and LA countries in 2020. The vertical black and grey lines report the median values for LA and OECD countries, respectively. Panel (a) shows that LA countries tax the rich at lower rates than the US, and the median TMTR in LA is only 30 per cent, which is significantly lower than the OECD median of 46.6 per cent. In fact, Bolivia, Paraguay, and Guatemala have TMTRs below 15 per cent. Panel (b) shows that the PIT system in LA excludes a much larger proportion of people than in the US and other OECD countries. The median exemption-to-GNIpc ratio in LA is six to seven times larger than in the US and the OECD median. The high filing and exemption thresholds exclude most formal workers from the PIT, much like the early PIT regimes of many rich countries (Jensen, 2022). As a result, the PIT in LA applies only to individuals at the very top of the income distribution, who pay a relatively large portion of the total PIT revenue. These features erode the PIT tax base and weaken the redistributive capacity of the tax system.

Figure 4:

Top marginal income tax rates and exemptions thresholds. (a) Top Marginal Tax Rates in 2020. (b) Ratio Exemptions to GNIpc in 2020.

Notes: This figure shows the 2020 top marginal income tax rates for the combined central and sub-central governments (panel (a)) and the exemption floor upon which people become liable relative to gross national income per capita (panel (b)). The black vertical line displays the median value for LA countries and the grey dashed vertical line corresponds to the median value for OECD countries. For each country, the exemption thresholds correspond to non-married individuals without dependants. In practice, individuals could earn more than the exemption floor and still pay no income tax by claiming family allowances (e.g. children, spouse, disability) or additional deductions (e.g. social security contributions, medical expenses, etc.). Four countries are not included in panel (b) due to data limitations: Venezuela, Turkey, Netherlands, and New Zealand. The OECD median excludes LA members: Chile, Colombia, Mexico, and Costa Rica.

Source: Own elaboration based on EY Worldwide Personal Tax and Immigration Guide for 2020 and World Bank national accounts data.

Low TMTRs on capital gains and wealth transfers

Figure 5 compares the TMTRs for capital gains tax (panel (a)) and inheritance and gift taxes (panel (b)) in LA and OECD countries. Most Latin American countries tax capital gains at a lower rate than OECD countries, with only a few exceptions. The median TMTR for capital gains in LA is less than half of the median rate in OECD countries. Additionally, the TMTR for inheritance and gift taxes are generally low in Latin America, especially when compared to the US.

Figure 5:

Capital gains and inheritance top marginal tax rates. (a) Capital Gains Tax. (b) Inheritance and Gift Tax.

Notes: This figure shows the top marginal tax rates for the capital gains tax (panel (a)) and the inheritance and gift tax (panel (b)). Tax rates for the inheritance tax correspond to those when the beneficiary is the spouse and children. The black vertical line displays the median value for LA countries and the grey dashed vertical line corresponds to the median value for OECD countries. The OECD median excludes LA members: Chile, Colombia, Mexico, and Costa Rica.

Source: Own elaboration based on PwC Worldwide Tax Summaries.

Self-employment at the top

Self-employment tends to be concentrated at the bottom and top of the earnings distribution in both developed and developing countries (Herrenõ and Ocampo, 2023). In particular, Latin Americans in the top decile of the earnings distribution—including doctors, lawyers, and other professionals—are more likely to be self-employed than individuals in the middle of the distribution. Self-employment income generally benefits from more generous tax allowances or is levied under tax-advantageous simplified regimes. Since it is generally self-reported, there is a large scope for evasion and underreporting (Kleven et al., 2011). Partly for this reason, in many developing countries the PIT amounts to little more than withholding taxes on formal wage earners (Bird and Zolt, 2005), hindering revenue collection and tax progressivity.

Work and business informality

Expanding the PIT base also requires tackling informality, which was widespread even before the Covid-19 pandemic (Maurizio, 2021). Informality, defined as working without contributing to social security, accounted for 70 per cent of all employment in developing and emerging economies, compared with about 18 per cent in developed countries (OECD and ILO, 2019). While the general view was that informality is more prevalent at the bottom than at the top of the earnings distribution (Busso et al., 2021), recent evidence suggests that undeclared (‘envelope’) wages at the top are substantial in numerous Latin American countries (e.g. Bérgolo and Cruces, 2014; Kumler et al., 2020; Bérgolo et al., 2021). For instance, in Brazil, 26 per cent of formal employees admitted to receiving compensation off-the-books equivalent to 22 per cent of their wage earnings (Feinmann et al., 2022). Crucially, the share of unreported wages increases with income, reaching 30 per cent for the top 5 per cent, which compounds the PIT loss and weak tax progressivity.

Furthermore, 81 per cent of all enterprises are informal globally (OECD and ILO, 2019), and most microenterprises are unregistered and do not file taxes (Bruhn and McKenzie, 2014; Brockmeyer et al., 2019).⁴ Although informality and selective non-filing are more common at the lower or middle part of the income distribution (Ulyssea, 2020), they create horizontal inequities between taxpayers, shift the tax burden to those who cannot escape scrutiny, and foster a culture of informality that erodes tax morale.

Tax allowances and evasion

There are many tax allowances in LA that further reduce the amount of income subject to the PIT base, which means that more income goes untaxed compared to the OECD. Specifically, for unmarried individuals, 54 per cent of average gross wage-earnings are untaxed in LA, while this fraction is only 14 per cent in the OECD (Barreix et al., 2017). Moreover, some LA countries have various non-standard tax reliefs for personal expenses like education, health, food, clothing, and housing. For instance, in Ecuador, these tax allowances could be as much as 50 per cent of an individual’s income (Bohne and Nimczik, 2018). Some of these tax allowances generate inequalities in the tax system. For example, pension income is largely untaxed in Colombia and this subsidy is regressive because pensions are larger for high-income taxpayers, and very few members of low-income households possess pensions. There is little evidence to support the justification of these tax allowances.

Simplified tax regimes (STRs), which aim to facilitate tax compliance for small businesses and the self-employed, are also prone to abuse by the rich and reduce tax progressivity.⁵ Recent evidence by Agostini et al. (2018) suggests that the rich in Chile use family businesses and STRs to lower their PIT liability and that over one-third of Chile’s richest 0.1 per cent owned at least one STR compared to only 2.6 per cent of taxpayers in the bottom 90 per cent.

Despite the lack of systematic cross-country evidence on the size of the PIT non-compliance, some studies suggest that LA countries have high evasion rates (ECLAC, 2020). For instance, using PIT declarations and third-party reports, Bérgolo et al. (2020) find that 15.7 per cent of Uruguayan employees underreport wages by 17 per cent of their tax liability, which is much higher than in developed countries like the US (Johns and Slemrod, 2010) and Denmark (Kleven et al., 2011). This situation affects the capacity of LA tax systems to raise revenue progressively, particularly since the PIT applies to the upper end of the income distribution, and the highest incomes account for a large share of all PIT revenues (IMF, 2015).

Wealthy Latin Americans often evade taxes by hiding income and assets overseas, with some countries reporting particularly high amounts of offshore wealth. Alstadsæter et al. (2018) estimate that the equivalent of 13 per cent of all Latin America’s GDP is held overseas, above the world average of 9.8 per cent, and this share increases to 25.7 per cent for Brazil, 36.5 per cent for Argentina, and 64.5 per cent for Venezuela. Many beneficial owners in the leaked Pandora Papers were from LA, especially Argentina, Brazil, and Venezuela (ICIJ, 2021). Indeed, the rich in Latin America are particularly prone to offshore evasion. For instance, Brounstein (2022) finds that tax haven use by Ecuadorians is highly skewed towards the top 0.1 per cent of the earnings distribution. Moreover, two in five Colombians in the top 0.01 per cent of the wealthiest distribution confessed to evading by hiding wealth—threefold the rate in Scandinavia and fourfold the rate in the Netherlands (Alstadsæter et al., 2019; Leenders et al., 2021; Londoño-Vélez and Ávila-Mahecha, 2021).

Administrative capacity

The effectiveness of state institutions is crucial for economic development, and efficient tax administration is necessary for providing public goods and services. In developing countries, where the need for investment in basic public goods and services is greater, ‘tax administration is tax policy’ (de Jantscher Casanegra et al., 1990).

Developing countries have significantly less tax capacity than rich countries. Figure 6 compares two measures of tax administration—operating expenditures relative to GDP and audit rates for the PIT—across LA countries, the US, and other OECD countries. Panel (a) shows that operating expenditures are 55 per cent larger in the median OECD country than in the median LA country (0.17 per cent compared to 0.11 per cent). However, there is significant variation across LA countries. Panel (b) shows that the PIT audit rate is three times higher in the OECD than in Latin America. Additional statistics from CIAT et al. (2022) reveal that investing in tax administration capacity could improve governments’ ability to tax Latin America’s rich. Only 59 per cent of LA countries use pre-filled PIT returns or assessments compared to 88 per cent across OECD countries, and only 35 per cent of LA countries have separate programmes for high-net-worth taxpayers, compared to 44 per cent in OECD countries.

Figure 6:

Tax administration expenditures and audit rates. (a) Operating Expenditures. (b) Audit Rates.

Notes: This figure shows the tax administration operating expenditures as a percentage of GDP (panel (a)) and audit rates (panel (b)) in fiscal year 2019 (or 2018 where 2019 data were not available). Operating expenditures refer to the overall level of resources devoted to tax administration including staff costs, ICT, and capital expenditure. We define the audit rate as the number of audited taxpayers divided by the number of active PIT taxpayers. This includes all audits and verification actions undertaken (excluding electronic compliance checks) regardless of whether tax adjustments were made or not. Due to data limitations we exclude Ecuador and El Salvador from panel (a) and Bolivia and the Dominican Republic from panel (b). The OECD median excludes LA members: Chile, Colombia, Mexico, and Costa Rica.

Source: Own elaboration based on ISORA 2020 survey (CIAT et al., 2022).

Tax morale

Income and wealth inequality are closely related to perceptions of fairness and social justice, which can have important implications for redistributive policy. Moreover, countries with higher levels of taxation as a percentage of GDP tend to have higher levels of tax morale, but tax morale has decreased in Latin America (OECD, 2019). This decline has been linked to the economic slowdown, poverty reversal, inequality, and corruption scandals across the region (OECD, 2018).

Figure 7 compares tax morale (panel (a)) and perceptions of corruption in the public sector (panel (b)) across selected countries around 2020. LA underperforms compared to OECD countries in both dimensions. For example, about twice as many Latin Americans believe that evading is justifiable compared to the OECD median. Similarly, the corruption perception index in Latin America is more than twice as high as in the OECD. If government institutions are perceived as being dishonest and corrupt and citizens are dissatisfied with the quality of public goods, this might hinder governments’ ability to collect revenue from direct taxes (Freytes, 2020). For instance, it is often said that Latin American elites are unwilling to pay taxes to provide public services for the masses because they can generally provide their own public services privately, ranging from public safety to education (Bird and Zolt, 2005).

Figure 7:

Tax morale and perceptions of corruption (circa 2020). (a) Evading is Justifiable. (b) Perceptions of Corruption.

Notes: This figure shows the degree of tax morale (panel (a)) and perceptions of corruption (panel (b)) for the US and a sample of LA countries. To measure low tax morale, we use the question from the World Values Survey about ‘Cheating on tax if you have the chance’ which ranges from 0 (never justifiable) to 10 (always justifiable). We calculate the proportion of the respondents who chose 5 to 10. The Corruption Perceptions Index (CPI) ranks 180 countries by their perceived levels of public-sector corruption, according to experts and business people. The index ranges from 100 (very clean) to 0 (highly corrupt). We plot the 100-index so that higher values correspond to higher perceived corruption. The black vertical line displays the median value for LA countries and the grey dashed vertical line corresponds to the median value for OECD countries. The OECD median excludes LA members: Chile, Colombia, Mexico, and Costa Rica.

Source: Own elaboration using data from the World Values Survey Wave 7: 2017–2022 and the 2019 Corruption Perceptions Index from Transparency International.

IV. Who are the rich and how much are they taxed?

One important aspect of the public policy debate about the top tail of the income and wealth distributions is the progressivity of the tax system at the top (Piketty et al., 2020). Evaluating rich individuals’ fiscal capacity requires measuring these upper tails well, often requiring combining household surveys with tax and social security records.⁶ Indeed, measuring inequality using tax data is crucial because capital incomes are less well covered in household surveys (Alvaredo et al., 2022). Recent evidence suggests that surveys’ ability to capture capital income has worsened in LA over time (Flores et al., 2022). As a result, tax and survey market incomes increasingly diverge for top fractiles in both levels and trends, making it difficult to measure inequality changes accurately. Notably, survey-based inequality estimates report an inequality drop in LA since the early 2000s (Clifton et al., 2020), but tax-based inequality estimates suggest that inequality has stagnated or increased (Alvaredo et al., 2022; Burdín et al., 2022; WID, 2022).

Unfortunately, progress in making tax records available to researchers in developing countries has been slow, and tax returns have limitations of their own when measuring inequality. The definition of ‘income’ often varies over time and across countries and can exclude important income components (Piketty et al., 2020). Moreover, as discussed before, only a small fraction of the population in developing countries files a tax return (e.g. Argentina, Brazil, Colombia, Peru), making it impossible to study tax-based inequality changes in the lower and middle part of the distribution. Tax evasion and tax avoidance may also vary over time and across countries depending on the strength of tax enforcement and the tax sheltering opportunities available.⁷ As a result, the study of rich or high-net-worth individuals in developing countries remains in its infancy.

(i). How rich are they?

Figure 8 combines survey and tax data for 2021 and compares the income needed to belong to the richest 1 per cent and top 0.01 per cent across some LA countries and three OECD countries: the US, France, and Spain. Panel (a) compares the top 1 per cent income threshold and shows that the income needed to belong to the richest 1 per cent varies substantially across countries in LA. Four groups of countries are identified, based on their top 1 per cent income threshold. At the bottom stands Ecuador, and three countries have thresholds of PPP USD 164,000–190,000 (Peru, Colombia, and Argentina). A third group comprises three countries with income thresholds ranging between PPP USD 200,000 and PPP USD 215,000 (Uruguay, Brazil, and Costa Rica), comparable with France and Spain. Indeed, despite these LA countries having substantially lower GDP per capita, rich individuals are comparable with France and Spain because income is highly concentrated in these countries. Lastly, Chile and Mexico stand out with the highest top 1 per cent income threshold in LA of around PPP USD 310,000, threefold Ecuador’s, and higher than all other LA countries, France, and Spain. The only country with a higher top 1 per cent income threshold is the US, and Mexico’s and Chile’s threshold is around two-thirds of that.

Figure 8:

The top income thresholds. (a) Top 1%. (b) Top 0.01%.

Notes: This figure shows the top 1 per cent income threshold across selected Latin American and more developed countries for 2021. Pre-tax national income is the sum of all pre-tax personal income flows accruing to the owners of the production factors, labour and capital, before taking into account the operation of the tax/transfer system, but after taking into account the operation of the pension system. The central difference between personal factor income and pre-tax income is the treatment of pensions, which are counted on a contribution basis by factor income and on a distribution basis by pre-tax income. The base unit is the individual (rather than the household), but resources are split equally within couples. The population is comprised of individuals over the age of 20.

Source: Own elaboration based on World Inequality Database (https://wid.world/). For methodological details, see https://wid.world/methodology/.

Furthermore, Panel (b) shows that the gap between the US and LA countries widens the higher one climbs the income ladder. For instance, the income at the 99.99th percentile in the US is more than twice the income in Chile and ninefold the income in Argentina, the country with the lowest income at the 99.99th percentile in the LA region.

(ii). Income composition and effective tax rates

The information available on the income composition and effective tax rates of top percentiles is limited. Unlike in the US, where many top earners are predominantly human-capital rich rather than financial-capital rich (Smith et al., 2019), Alvaredo and Londoño-Vélez (2013) find that high-income Colombians own capital stock and capital income makes up the largest income share at the very top of the distribution. Because capital income is taxed favourably, as in many other countries worldwide, personal income taxation has little effect on inequality, and the system becomes regressive at the very top. Indeed, Colombian top incomes face statutory TMTRs comparable to OECD countries, but legal erosion of the tax base reduces the effective tax rate, defined as tax liability relative to total income, to 7–8 per cent for the top 1 per cent and less than 4 per cent for the top 0.01 per cent of the income distribution.

Similar findings are consistent in other LA countries, with capital being the predominant source of income at the top (Cano, 2018; AFIP, 2020; Burdín et al., 2022). Effective tax rates often fall within the top 1 per cent due to preferential tax treatment of capital income, and high-income individuals are likely to further reduce their taxable income through legal tax deductions and exemptions. In Ecuador, the top 1 per cent faced an average effective tax rate of about 7 per cent, while the richest 0.001 per cent paid less than 2 per cent in personal income taxes.

Business ownership is another important source of income for the rich (Smith et al., 2019; Kopczuk and Zwick, 2020), but most large businesses are taxed separately by the corporate tax in most countries. Since corporate ownership of individuals is generally not recorded, individual owners are rarely identified (Piketty et al., 2020). Moreover, ownership of partnerships and close corporations is also opaque. This introduces measurement error in top-income shares and has severe implications for tax progressivity. Additionally, corporations are generally taxed at lower rates than top personal incomes, providing business owners with incentives to leave profits in the firm, and independent professionals prefer to incorporate and consume profits without declaring dividends to avoid personal income taxation. For example, luxury vehicles for personal use may be registered to the firm and distributed profits may be omitted from tax returns.

Recent evidence from Chile by Fairfield and Jorratt De Luis (2016) sheds light on this phenomenon. Unlike in most countries, Chilean corporate ownership information is systematically collected, allowing researchers to link corporate wealth and profits to final individual owners and re-estimate top income shares when accounting for accrued and undeclared distributed profits. Adjusting for undeclared distributed profits raises the top 1 per cent income share from 15 per cent to 22–26 per cent, and including accrued profits instead of distributed profits further raises the top 1 per cent share to 32–33 per cent. Despite this income concentration, the top 1 per cent pays modest average effective income tax rates of 15–16 per cent, both with and without including the corporate tax in the numerator and accrued profits in the denominator.

V. Recurrent taxes on wealth

Latin America has a long tradition of taxing the wealth of individuals and firms. Historically, countries have taxed wealth because it may be easier to observe assets, such as land and housing, than income. Indeed, wealth taxes in Latin America were generally not brought about by a leftist movement for progressive redistribution to curtail inequality, but rather by concerns about income tax evasion and the need to secure revenue for public goods.

Given the political ‘impossibility’ of enforcing taxes on Latin America’s rich and powerful taxpayers (Bird, 1992), why did the LA elites support wealth taxation? Freytes (2020) argues that Latin American elites have supported wealth taxes when they saw the consolidation of a capable central state as advancing their own interests. For example, Colombian elites supported reintroducing the wealth tax when a centre-right-wing government in 2002 proposed a ‘temporary’ wealth tax to finance the exigencies of war against illegal armed groups. The combination of fiscal and security crises, coupled with improving perceptions of the government’s provision of public safety, generated cohesion among Colombian business and government elites in favour of a wealth tax (Flores-Macías, 2014). Similarly, Argentina successfully implemented a tax amnesty only when the elites supported President Macri’s policy. Both Colombia and Argentina made use of earmarked tax revenues to get popular support for the new wealth tax and tax amnesty, respectively. Thus, Latin America’s experiences with wealth taxation and enforcement suggest that increasing state capacity requires, in addition to better enforcement, overcoming opposition from business and economic elites and leveraging political opportunity.

As with most other wealth-taxing countries, the wealth tax is not a significant source of revenue in Latin America and rarely exceeds 0.1 per cent of GDP (OECD, 2018; Jorratt, 2021). This is partly because of their high exemption thresholds and widespread evasion, as enforcing wealth taxation is notoriously challenging (Kopczuk, 2019; Advani and Tarrant, 2021; Scheuer and Slemrod, 2021). Generally, it is hard to identify assets owned by taxpayers. While governments can use third-party information reported by banks and other financial institutions on clients’ financial assets, like bank statements, other assets, like unincorporated business assets, are not third-party reported and can be easily hidden (Londoño-Vélez and Ávila-Mahecha, 2022). Moreover, foreign assets, which are predominantly held by the wealthiest individuals (Alstadsæter et al., 2019; Londoño-Vélez and Ávila-Mahecha, 2021; Londoño-Vélez and Tortarolo, 2022), have historically been hard to observe. While recent automatic tax information exchange agreements (TIEAs) can help shed light on offshore financial accounts, foreign real estate remains excluded from these multilateral agreements.

Since many wealthy Latin Americans own foreign real estate (Londoño-Vélez and Ávila-Mahecha, 2021; Londoño-Vélez and Tortarolo, 2022), authorities are still unable to observe a significant portion of foreign wealth. Lastly, it is hard to value some assets if they are not reported in tax records at market values (e.g. real estate) or if they are inherently hard to value (e.g. stocks in closely held private businesses).

These difficulties faced by wealth-taxing countries can be more severe in the developing world. For example, property taxes are generally based on outdated cadastral values and property transactions are rarely used to update cadastral values (Sepulveda and Martínez-Vázquez, 2012). Moreover, the value of other assets is based on historical values that are not aligned with the market value of the asset. In addition, resources for tax audits can be meagre and there are no dedicated units for managing the affairs of high-net-wealth taxpayers. Furthermore, assets can be severely underdeclared due to the pervasiveness of the ‘shadow’ (underground, informal) economy. Lastly, there can be considerable self-reporting and avoidance channels due to the aforementioned enforcement challenges in developing countries.

Crucially, however, these enforcement challenges are not unique to wealth taxation, as there are many challenges in taxing wealth transfers, capital income, and real property. For example, the estate and gift taxes in the US collect little revenue due to tax loopholes and high exemption thresholds. In addition, numerous studies have shown that a lack of verifiable paper trails also makes income and value-added taxes difficult to enforce (Kleven et al., 2011; Pomeranz, 2015). Furthermore, weak cadastres (property registers) and outdated cadastral values erode the property tax base in many countries (Sepulveda and Martínez-Vázquez, 2012; Brockmeyer et al., 2022). Lastly, tax havens pose a global threat to capital income taxation (Zucman, 2013; Tørsløv et al., 2022).

Despite the weaknesses, it may be surprising to learn that several LA countries still impose wealth taxes. While rich countries progressively gave up on taxing wealth (OECD, 2018; Saez and Zucman, 2019), Argentina, Colombia, and Uruguay currently levy annual wealth taxes (ECLAC, 2021), and Bolivia and Ecuador have implemented temporary wealth taxes in the Covid pandemic aftermath. Moreover, several countries, including Brazil, require all income taxpayers to report their assets and liabilities annually to support income tax enforcement, as mandatory wealth reporting enables authorities to check if a change in the taxpayer’s net worth is compatible with a change in her reported income.^8,9

Why do wealth taxes still exist in Latin America? First, wealth taxes are used to complement the ineffective taxation of personal income, bequests and gift taxes, capital income, and property (OECD, 2022b, OECD et al., 2022). Second, the high levels of income and wealth inequality in the region make it possible to generate revenue by taxing wealth. Third, there is generally political support for wealth taxes in Latin America, particularly in the aftermath of the pandemic. Fourth, many Latin American countries are part of the OECD’s Common Reporting Standard for the Automatic Exchange of Information on financial accounts between tax authorities, which increases the risk (or at least the perceived risk) of detecting offshore tax evasion. Lastly, the next section demonstrates how strengthening the enforcement regime can increase reported wealth and improve tax progressivity, even in settings with low baseline tax compliance.

VI. How the rich respond to taxation and enforcement

The existing evidence on how the rich respond to taxes and enforcement predominantly focuses on rich countries (Auten et al., 2016; Advani and Tarrant, 2021; Scheuer and Slemrod, 2021), and much less is known in developing countries. This section briefly discusses the evidence of how the rich in LA respond to progressive taxes on income and wealth and changes in enforcement policy.

(i). Responses to wealth taxes, havens outflows taxes, and inheritance and gift taxes

Londoño-Vélez and Ávila-Mahecha (2022) study how taxpayers respond to Colombia’s personal wealth tax using tax microdata on individual wealth holdings. The authors link the tax records with the leaked Panama Papers, which shed light on offshoring to Colombia’s most relevant tax havens. For identification, they leverage cross-sectional and time tax variation from four reforms that substantially changed the wealth tax design using bunching and difference-in-difference techniques. They find that wealth tax hikes cause taxpayers to instantly lower their reported wealth to fall below the tax bracket cut-offs. Crucially, however, these effects can persist for years—even after the wealth tax is no longer in place. This occurs because taxpayers behave strategically to avoid detection and future wealth taxes. As a result, transitory tax policy can have a hysteresis effect. Several pieces of evidence suggest that tax sheltering drives taxpayers’ responses. Specifically, taxpayers misreport what tax authorities cannot cross-verify, such as (underreporting) business assets and inflating debts. In addition, wealthy individuals avoid wealth taxation by hiding assets in hard-to-track entities in tax havens.

Can tax policy be used to discourage offshore tax avoidance? In addition to measures such as multinational tax coordination, information-sharing agreements, and voluntary disclosure programmes, governments can implement a fiscal havens outflows tax to increase the cost of transacting with tax havens. A study by Brounstein (2022) examines the effects of such a tax policy in Ecuador. The study finds that an increase in the outflows tax for dividends sent to tax havens resulted in a reduction of dividend outflows to tax havens, with a net-of-tax elasticity of dividend payments abroad ranging from 13 to 40. Additionally, individuals subject to the policy reported 40 per cent more domestic income and paid 55 per cent more in income taxes due to increased reporting of capital income and independent labour income flows. Later, we review other measures that have been taken to combat offshore tax avoidance.

The study by Locks (2023) examines how people in Brazil respond to inheritance and gift taxes. In contrast to the US, where very few taxpayers are subject to these taxes due to high exemption thresholds, Brazil’s inheritance and gift taxes affect more ordinary citizens. The author looks at tax data from 2014 to 2020 and uses tax reforms and differences in tax rates across regions of Brazil to identify the effects of tax changes on wealth transfers. The study finds that Brazilians react strongly to tax increases, and they time their wealth transfers in ways that minimize their tax liability. Specifically, the short-term elasticity of gift transfers with respect to the net-of-tax rate is 20. In addition, the author uses a difference-in-differences approach to estimate medium-term effects and finds that tax evasion increased, with more people in the bottom and middle of the wealth distribution failing to report their estates.

(ii). Responses to income taxes

Bérgolo et al. (2022) study the effects of a tax reform that increased the top marginal tax rate on labour income in Uruguay. This reform also amplified the tax rate differential between this tax and other taxes that levied earnings of top income earners. Using administrative tax records, the authors find that top earners significantly reduce their reported labour income, with the strongest response coming from those at the very top of the earnings distribution.

The intensive margin elasticity is close to 0.6, consistent with the average income elasticity before deductions reported in a meta-regression study by Neisser (2021). The authors also show a strong extensive margin reaction to the tax increase: a 1 per cent reduction in the average net-of-tax rate leads to a 2.5 percentage point decline in the probability of reporting earnings. Wage earners vanished from the tax records, while self-employed individuals shifted their labour income to other tax regimes, namely, the corporate tax base, as Kopczuk and Zwick (2020) documented in the US.

Across the border, Argentina gave a temporary income tax holiday to high-wage earners. A study by Tortarolo et al. (2020) examines how taxpayers’ labour supply responds to this tax holiday. The authors compare workers earning above and below the eligibility rule using Social Security data. Perhaps surprisingly, the authors find that tax-exempt high-income workers do not substantially respond to the tax holiday: the wage earnings elasticity is 0.017. This diminished response is primarily due to flexible earnings components, such as overtime hours. Tax-exempt workers, such as job switchers and executives, respond more to the tax holiday, as do new entrants to the labour market. The authors argue that rigidities in the labour market prevent wage earners from responding forcefully to tax changes, as others have found in more developed countries (e.g. Cahuc and Carcillo, 2014).

(iii). Tax compliance at the top

Tax evasion is a long-standing policy concern, and enforcing taxes is key for developing a functioning state and economy (Besley and Persson, 2009, 2013; Pomeranz and Vila-Belda, 2019; Slemrod, 2019). Moreover, effectively taxing the elites is key for redistribution. Recent research indicates that altering enforcement policies can deter rich individuals from utilizing complex tax avoidance tactics like simplified tax regimes (STRs) for business owners or offshore tax evasion.

Tax codes often allow business owners and self-employed taxpayers to choose between personal or corporate bases for tax purposes, and several researchers have shown that high-income earners shift tax bases to avoid taxation. Chile sought to tackle this by restricting the opportunities for tax planning through STRs, and Agostini et al. (2018) show that led to a decrease of 10–15 per cent in reported individual incomes for businesses using that regime. In contrast, earnings reported from alternative sources increased, resulting in an overall increase in the taxable income of 4–7 per cent. Since rich individuals are particularly prone to using STRs, closing tax loopholes by making STRs less tax advantageous expanded the tax base and improved tax progressivity.

In addition to aggressive tax planning, rich individuals evade by offshoring income and wealth to tax havens. The pervasiveness of offshore evasion has prompted many countries to conduct a series of enforcement initiatives to improve wealthy taxpayers’ tax compliance, including the OECD’s Common Reporting Standards and cross-border automatic TIEAs to trace taxpayers’ foreign income and assets. Moreover, many countries have implemented voluntary disclosure programmes or tax amnesties to entice wealthy tax evaders to disclose their foreign incomes and assets in exchange for reduced penalties and no prosecution (OECD, 2015). Two papers suggest that these enforcement policies have contributed to effectively boosting tax revenue and improving tax compliance at the top.

Londoño-Vélez and Ávila-Mahecha (2021) examined a Colombian government programme that encouraged wealthy individuals to disclose their hidden wealth between 2015 and 2017. The programme was highly successful, recovering hidden wealth equivalent to 1.73 per cent of the country’s GDP, which is significantly more than what a similar programme in the US recovered (Johannesen et al., 2020). The authors analysed the impact of the programme’s tax incentives on income and wealth tax compliance over time using a difference-in-differences approach. The results show that the policy significantly improved tax compliance, as 3 years after the disclosures, disclosers reported 49 per cent more wealth. By virtue of disclosing the return of those assets, disclosers paid almost 40 per cent more income taxes. Importantly, since wealth tax evasion is most prevalent among the wealthiest individuals, the programme effectively raised revenue from this group, making the tax system more progressive.

Argentina has also sought to crack down on offshore tax evasion to boost tax collection and social spending. Despite previous unsuccessful attempts, a 2016 amnesty was successful, with wealthy Argentines disclosing assets worth 21 per cent of the country’s GDP. Londoño-Vélez and Tortarolo (2022) found that the amnesty, in combination with automatic TIEAs, led to an increase in the reported assets by the wealthiest 0.1 per cent of adults. Four years after the amnesty, this group reported owning two to three times more assets than before the programme. As a result of these disclosures, Argentines today report that nearly one-half of all their assets are offshore, aligning more closely with macro estimates of offshore wealth (Alstadsæter et al., 2018). Moreover, the wealth and capital income tax bases more than doubled. In particular, the expanded wealth tax base allowed authorities to collect more revenue by subsequently taxing foreign assets at very high rates.

The findings of these studies in two Latin American countries suggest that enforcing tax laws can make it possible to have progressive taxation on capital in today’s globalized world. This is true even in situations where there is high inequality and many people do not comply with tax laws.

VII. Policy discussion

To effectively tax rich individuals, it is necessary to broaden the tax base by including capital income, foreign income, and sole proprietorships and partnership income. In this section, we provide insights on policy choices based on lessons drawn from the paper.

First, LA governments should re-evaluate their standard practices of taxing capital income and capital gains at preferential rates. These practices create opportunities for arbitrage that disproportionately benefit wealthy individuals. Recent studies in the US show that lowering the capital gains tax rate does not pay for itself and increasing this rate could have a significant impact on revenue and tax progressivity (Agersnap and Zidar, 2021).

Second, tax authorities should systematically and rigorously monitor foreign income and assets. Non-compliant taxpayers should be given the chance to voluntarily disclose hidden income and wealth following best practices to safe-keep the tax morale of tax-compliant taxpayers (OECD, 2015). Tax authorities should leverage information from automatic TIEAs to improve tax enforcement. Governments should strive for better international coordination to ensure that foreign capital income and offshore wealth are taxed fairly. Tax administrations should also establish dedicated units to scrutinize high-net-worth taxpayers to ensure proper taxation.

Third, special attention should be given to the tax treatment of sole proprietorships and partnership income to prevent abuse. For example, governments should take steps to address the misuse of STRs by rich individuals.

More generally, LA governments should enhance the role of the PIT since most people in the region do not pay income taxes and rich individuals pay low effective tax rates on their capital, personal business, and labour income. It is true that TMTRs in LA are lower than in developed economies, but it is unclear whether governments should increase them since the low rates could be due to the region’s weak enforcement capacity and rampant avoidance and evasion opportunities. Instead, the first step should be to broaden the PIT base. To achieve this, governments should reduce the number of exempt and non-taxable income items within the PIT. Additionally, policy-makers should reconsider non-standard tax reliefs from personal expenses and other tax allowances that primarily benefit top-income taxpayers. Governments should systematically evaluate the cost–benefit of tax allowances and expenditures to estimate the revenue forgone from each item and determine whether they are desirable policy instruments. For instance, some LA governments, such as Colombia, have commissioned in-depth assessments of their tax expenditure reporting practices (OECD, 2022a), which is a welcome step forward.

To further broaden the PIT tax base, policy-makers will need to consider lowering the high statutory exemption thresholds, perhaps establishing low tax rates for the new (and lower-income) taxpayers. Although exemption thresholds have received little attention in the academic literature, they are an important policy tool for collecting taxes and redistributing income (Jensen, 2022). Additionally, recent research suggests that expanding the tax base by including more taxpayers can have a ‘participation dividend’, promoting inclusive governance and stimulating political engagement (Gadenne, 2017, Weigel, 2020, Martinez, 2023). To successfully implement tax-broadening policies, it is important to couple the lowering of exemption thresholds with an expansion of the coverage and use of third-party reported information. Increasing the amount of employee information trails by employers can reduce the administrative costs of broadening the tax base.

In general, to enhance the role of the PIT and effectively tax the rich, governments must improve tax administration capacity. This includes investing in technology that can leverage third-party data to pre-populate PIT returns, facilitate electronic or return-free filing, and detect and deter tax evasion using modern data science techniques. Governments should also increase resources to conduct audits and use modern technology to improve targeting efficiency. To prosecute tax evaders effectively, courts and law enforcement institutions must be strengthened.

As LA governments work to streamline their income tax systems, wealth taxes can serve as a complementary tool to boost the effective tax rate at the top of the wealth distribution. However, wealth taxes could be improved by updating property registers to reflect market values and, as discussed above, scrutinizing foreign assets more closely.

Lastly, to evaluate rich individuals’ tax capacity and draw lessons on how to improve the design of tax policy, researchers should be given better and more transparent access to tax data. Collaborating more deeply between tax authorities and researchers is essential to produce high-quality research that can guide the policy debate and strengthen the tax system.

Appendix

Table 1:

List of countries by world region

Latin America	Europe	Oceania	North America	Africa	South/East Asia	West Asia /Central/East
Argentina*†	Albania	American Samoa	Bermuda	Algeria	Afghanistan	Armenia†
Brazil*†	Andorra	Australia*†	Canada*†	Angola†	Bangladesh	Azerbaijan
Chile*	Austria*†	Cook Islands	Greenland	Benin	Bhutan*	Bahrain
Colombia*†	Belgium*†	Fiji*	Saint Pierre and Miquelon	Botswana*	Brunei Darussalam	Belarus†
Costa Rica*†	Bosnia and Herzegovina	French Polynesia	USA*†	Burkina Faso*	Cambodia*	China*†
Ecuador*	Bulgaria*	Guam		Burundi	India	Georgia
El Salvador*†	Channel Islands	Kiribati		Cabo Verde*	Indonesia*†	Hong Kong
Mexico*†	Croatia†	Marshall Islands		Cameroon*	Iran†	Iraq
Peru*†	Cyprus†	Micronesia		Central African Re- public	Lao PDR*†	Israel*†
Uruguay*†	Czech Republic*†	Nauru		Chad	Malaysia†	Japan*
	Czechoslovakia	New Caledonia		Comoros	Maldives*	Jordan
	Denmark*†	New Zealand*†		Congo	Myanmar	Kazakhstan*†
	Estonia*†	Niue		Cote d’Ivoire*	Nepal	Korea*†
	Faroe Islands	Northern Mariana Islands		Djibouti	Pakistan*†	Kuwait
	Finland*†	Palau		DR Congo*	Philippines*†	Kyrgyzstan*
	Former Czechoslo- vakia	Papua New Guinea		Egypt*	Singapore*†	Lebanon
	France*†	Samoa		Equatorial Guinea	Sri Lanka	Macao
	German Democratic Republic	Solomon Islands*		Eritrea	Thailand*†	Mongolia*†
	Germany*†	Tokelau		Ethiopia	Timor-Leste	North Korea
	Gibraltar	Tonga		Former Ethiopia	Viet Nam*†	Oman
	Greece*†	Tuvalu		Gabon		Palestine
	Holy See	Vanuatu		Gambia		Qatar
	Hungary*	Wallis and Futuna		Ghana		Russian Federation
	Iceland*			Guinea		Saudi Arabia
	Ireland*†			Guinea-Bissau		Syrian Arab Republic
	Isle of Man			Kenya*		Taiwan
	Italy*†			Lesotho		Tajikistan
	Kosovo			Liberia		Turkey*†
	Latvia*†			Libya		Turkmenistan
	Liechtenstein*			Madagascar*		Ukraine
	Lithuania*			Malawi		United Arab Emirates
	Luxembourg*†			Mali*		USSR
	Malta†			Mauritania*		Uzbekistan
	Moldova†			Mauritius*		Yemen
	Monaco			Morocco*
	Montenegro†			Mozambique
	Netherlands*†			Namibia*
	North Macedonia			Niger*
	Norway*†			Nigeria†
	Poland*†			Rwanda*
	Portugal*†			Saint Helena
	Romania†			Sao Tome and Principe
	San Marino			Senegal*
	Serbia†			Seychelles*†
	Slovakia*†			Sierra Leone†
	Slovenia*†			Somalia
	Spain*†			South Africa*
	Sweden*†			South Sudan
	Switzerland*†			Sudan
	United Kingdom*†			Swaziland
	Yugoslavia			Tanzania
				Togo
				Tunisia*
				Uganda
				Western Sahara
				Zambia
				Zanzibar
				Zimbabwe

Notes: This table lists countries grouped by world regions used to construct the Figures in section II. Symbol * indicates countries used in Figure 2; symbol ^† indicates countries used in Figure 3. All countries listed in this table were used to generate Figure 1.