Globalization has opened new tax evasion opportunities, such as owning undeclared assets abroad through offshore accounts, shell companies, and trusts. Moreover, wealthy individuals can leverage sophisticated legal advice and financial engineering to hide income and assets — legally, illegally, or in a grey zone. Guyton, Langetieg, Reck, Risch, and Zucman aim at better measuring tax evasion by wealthy individuals and at constructing revised estimates of top income and wealth shares factoring in tax evasion. To improve estimates of the size and distribution of tax evasion in the United States, the researchers combine National Research Program data with new data on American taxpayers’ holdings in offshore tax havens (leveraging, e.g., data from the Offshore Voluntary Disclosure Program (OVDP) and FBAR returns). Incorporating these new offshore data enables the researchers to provide a more accurate estimate of non-compliance at the top-end of the distribution of income and wealth. They use their estimates of the distribution of evasion to investigate the main covariates of tax evasion and develop tools to better predict non-compliance at the top. The ultimate goal is to make progress on estimating and understanding currently un-detected tax evasion, and its implications for the measurement of inequality.
Bryant, Koch Helsø, and Mitnik carry out a comparative analysis of inequality of opportunity for long-run income in Denmark and the United States. They use high-quality administrative data for both countries as well as samples that represent the full populations of interest. In addition, they rely on improved methods and advances a plausible identification assumption that allows to legitimately compare lower-bound estimates of inequality of opportunity across the two countries. There are three main findings. First, inequality of opportunity for long-run income is very high in the United States. With types based only on gender and parental income rank as circumstances outside of people's control, the lower-bound Gini coefficients for individual earnings and family income opportunities are around 0.24. An extension of the analysis to also account for race and ethnicity suggests an upward adjustment of the Gini coefficient for family income opportunities of 14 percent -- this implies that no less than 52 percent of long-run family income inequality is due to circumstances outside of people's control. Second, inequality of opportunity for long-run income is far from negligible in Denmark. The lower-bound Gini coefficients for individual earnings and family income opportunities are around 0.10. Lastly, inequality of opportunity for long-run income is radically higher in the United States than in Denmark, and this result is very robust to the inequality measure employed in the analysis. For long-run disposable income opportunities, which factor in taxes and public transfers (including refundable tax credits), the Gini coefficient is no less than 2.8 times higher in the United States whereas the mean logarithmic deviation is no less than 8.2 times higher.
It is widely recognised that household surveys do not fully capture the incomes of the very richest individuals and households, particularly those among the so-called “top 1%”, for reasons including non-response and under-reporting. As a consequence, estimates based on survey data alone typically understate true levels of inequality. Webber, Tonkin, and Shine present new research and analysis to develop a methodology for improving the measurement of the upper tail of the distribution, which is suitable for use in ONS’s official statistics on household income, in terms of being methodologically sound and based on robust academic research, transparent and understandable by users, and an approach where adjustments are made to underlying microdata rather than aggregates. The methods presented in the research build upon the work of both the UK Department for Work and Pensions and Burkhauser et al. (2018a) in employing methods in which survey-based mean incomes for quantile groups at the top of the distribution are replaced by equivalent figures from tax data. The analysis examines two sets of methods developed from these approaches, with variants of each tested to determine the most appropriate methodology to apply in future official statistical releases by ONS.
There is a clear interest in the development of measures of economic well-being and inequality that are based upon and consistent with the national accounts framework. Such an approach offers the potential for improved international comparability, greater coherence within countries, and opportunities for enhanced frequency and timeliness of distributional measures. In recent years, there has therefore been a growing body of work seeking to produce distributional national accounts, including that coming out of the OECD-Eurostat Expert Group on Disparities within a National Accounts Framework (EG DNA). However, in many countries, a key challenge in producing such measures is the coherence between micro and macro sources, with household surveys and national accounts aggregates differing, often substantially, from each other. There are two principal reasons for this. The first is differences in recorded amounts, which may reflect issues with survey coverage, non-response and underreporting, as well as measurement error in the national accounts. The second is definitional differences, reflecting the different purposes to which the two sources are traditionally put. Tonkin, White, Stoyanova, Youssef, Sidhu, and Payne build upon recent research into both sets of reasons, in order to develop new indicators of inequality, poverty and shared prosperity based upon and consistent with national accounts. Data from the United Kingdom is used as an example of how such indicators can be produced for a wider group of countries. Their analysis highlights how such indicators may differ from those based on survey microdata alone and how they may provide new and complementary insights, supporting the timely monitoring of inequalities and inclusive growth at both the national and international levels.
Using tax data from several Swiss cantons, Martinez presents a set of empirical facts regarding the distribution of wealth and income in Switzerland. First, they shed light on the composition of wealth and income along the distribution, including the very top. They find substantial heterogeneity in the composition for different population groups. Second, they document the joint distribution of income and wealth, where they find a strong correlation between the two, especially at the very top. Third, wealth mobility over a ten year period is substantially lower than income mobility. Persistence in wealth rank is especially strong in the tails: the bottom 20% seem to be stuck in a wealth trap, while those in the top 1% hardly ever leave the top 10% at all unless they die. Fourth, inter-vivos gifts and inheritances substantially increase intragenerational wealth mobility. At the same time, rich individuals are more likely to receive an inheritance and average amounts bequested rise with the wealth rank of the heir. This finding suggests a strong correlation between the wealth rank of those leaving and those receiving an inheritance.
Earlier work has established that the US has exceptionally high inequality of disposable household income (i.e., income after accounting for taxes and transfers). There is a debate whether it is due to an unusually high inequality of market (pre-tax-pretransfer) income or to weak redistribution. Gornick, Milanovic, and Johnson look more deeply at market income inequality, focusing on its main component -- labor income -- across a group of 24 OECD countries. They disaggregate the working-age population into household types, defined by the number and gender of the household's earners and the partnership and parenting status of its members. The researchers conclude that within-group inequality of labor incomes in the US is, in almost all groups, high by OECD standards. The roots of US inequality exceptionalism are not to be found in an unusual demographic composition, nor in unusually high or low mean incomes of some demographic groups, but in pervasive high inequality within each of these groups.
In recent years, “inequality” has received an extraordinary amount of attention in political, policy, and academic circles. In the U.S. the conversation has been overwhelmingly national in scope. This national focus misses another enormously consequential axis of American inequality, one that has received inadequate attention in contemporary academic and policy circles – that is inequality by geography, specifically inequality across the 50 U.S. states. In this paper Bruch, Gornick, and van der Naald contribute conceptually and empirically to our understanding of the role of subnational government (states) in social provision and redistribution, directing attention to the consequences of safety net decentralization—especially inequalities in social provision, and inequalities in poverty reduction. Using an unique dataset of comparable social provision indicators, we examine the magnitude of cross-state variation in the generosity of benefits and the inclusiveness of safety net provisions across the U.S. states, and compare it to the magnitude of cross-national policy variation among a set of high-income countries. The researchers find that there is substantial cross-state inequality in social provision, and that this level of variation rivals the variability that is observed cross-nationally. To examine redistribution, they focus on poverty reduction among working-age households with children using the CPS ASEC data to explore the role of four redistributive mechanisms: federal transfers, state transfers, federal taxes, and state taxes. Bruch, Gornick, and van der Naald find that state transfers and federal taxes reduce market income poverty rates substantially, and that the magnitude of these redistributive impacts varies across states. They conclude that these state-level variations in social provision and redistribution are meaningful forms of inequality – inequality in the treatment of similar needs and claims by people who happen to live in different states, and argue that this form of inequality deserves more sustained attention, particularly in regard to policy design.
Recent U.S. estimates of upward income mobility are based on theconditional expectation function of child income rank given parentincome rank. Asher, Novosad, and Rafkin use interval data methods to generate comparablestatistics for educational mobility for different racial groups in theUnited States. Aggregate intergenerational educational mobility isnearly identical to recent income mobility estimates, but there aresubstantial differences across racial groups. In particular, theblack-white gap for men is smaller for educational mobility than forincome mobility, but the gap is larger for women. Black women born toparents in the least educated 50% can expect to attain a rank fourpoints lower than white women born to similar parents.
Using data from the Census Bureau's Longitudinal Employer-Household Dynamics infrastructure files, Abowd, McKinney, and Sabelhaus study changes over time and across sub-national populations in the distribution of real labor earnings. While much is known about earnings inequality, mobility, and volatility at the national level, much less is known about earnings distributions and distributional dynamics at the sub-national level. The analysis of earnings dynamics is comprehensive, including labor market entry and exit and movements into and out of local labor markets. The researchers consider four large MSAs (Detroit, Los Angeles, New York, and San Francisco) for the period 1998 to 2017, with particular attention paid to the sub-periods before, during, and after the Great Recession. The results contribute to the emerging literature on differences between national and regional economic outcomes, and exemplify what will be possible with a new data exploration tool -- the Earnings and Mobility Statistics (EAMS) web application -- currently under development at the US Census Bureau.
Inequality in income, consumption, and wealth is increasing, and inequality in the joint distributions is increasing faster than inequality in any of the single distributions (Fisher, Johnson, Smeeding, and Thompson, 2018). Studying the joint distribution of income, consumption, and wealth tells us something about past well-being, current well-being, and future well-being. Fisher and Johnson use the Panel Study of Income Dynamics (PSID), which has followed individuals and families over almost five decades. The PSID has been the benchmark source for measuring both intra- and inter-generational mobility, and it is the only data set with income, consumption and wealth. The researchers' study builds on previous work (Fisher et al. (2016)) and extends these results back to 1968. Following the methods in Fisher and Johnson (2006), the researchers impute consumption to the earlier years in the PSID to obtain measures of inequality and mobility from 1968 to 2017. They find that consumption mobility is higher than income mobility, which is higher than wealth mobility. They also find that people with low wealth are less likely to move up relative to those with high wealth. By examining cohorts, the researchers find that inequality increases and mobility falls within each cohort and that the younger cohorts experience higher inequality and lower mobility.
The patterns and determinants of household wealth accumulation have long been of interest to economists. Using the 1989 and 2016 Surveys of Consumer Finance, Gale and Harris examine the determinants of changes in mean and median wealth of households of given ages across different birth cohorts. While the Great Recession reduced wealth in all age groups, longer-term trends show that wealth of older age groups has increased while the wealth of younger age groups has declined. The researchers show that a substantial share of these changes, in both directions, can be explained by the evolution of household demographic and economic characteristics.
Kukk, Meriküll, and Rõõm study the gender gap in net wealth. They use administrative data on wealth that are linked to the Estonian Household Finance and Consumption Survey, which provides individual-level wealth data for all household types. The researchers find that the unconditional gender gap in mean wealth is 45% and that it is caused by large wealth disparities in the upper end of the wealth distribution. The structure of assets owned by men is more diversified than that for women. Men own more business assets and vehicles, while women own more deposits. The gender gaps in these asset components cannot be explained by observable characteristics. For partner-headed households the raw gender gaps across deciles are mostly in favour of men, and more strongly so for married couples, indicating that resources are not entirely pooled within households. For single-member households the raw gaps across quantiles are partially in favour of women. Accounting for observable characteristics renders the unexplained parts of the gaps mostly insignificant for all household types.
Batty, Bricker, Briggs, Friedman, Nemschoff, Nielsen, Sommer, and Henriques Volz describe the construction of the Distributional Financial Accounts (DFA), a dataset containing quarterly estimates of the distribution of US household wealth since 1989. The DFA build on two existing Federal Reserve Board statistical products -- quarterly aggregate measures of household wealth from the Financial Accounts of the United States, and triennial wealth distribution measures from the Survey of Consumer Finances -- to incorporate distributional information into a national accounting framework. The DFA complement other sources by generating distributional statistics that are consistent with macro aggregates, by providing quarterly data on a timely basis, and by constructing wealth distributions across demographic characteristics. The researchers encourage policymakers, researchers, and other interested parties to use the DFA to better understand issues related to the distribution of US household wealth.
Wealth inequality in the US is high and rising, but Social Security is generally not included in those wealth measures. Social Security Wealth (SSW) is the present value of future benefits that an individual will receive less the present value of future taxes they will pay. When an individual enters the labor force, they generally face a lifetime of taxes to pay before they will receive any benefits, and thus their initial SSW is generally very low or negative. As an individual works, their SSW grows and generally peaks somewhere around typical Social Security benefit claim ages. The accrual of SSW over the working life is most important for lower-income workers, because the progressive Social Security benefit formula means that taxes paid while working are associated with proportionally higher benefits in retirement. Adding SSW to existing household wealth measures, which Sabelhaus and Henriques Volz do here using the Survey of Consumer Finances (SCF) for 1995 through 2016, provides a more comprehensive view of wealth inequality and wealth profiles over the lifecycle, across wealth groups, and over time.
In addition to the conference paper, the research was distributed as NBER Working Paper w27110, which may be a more recent version.
Garbinti and Savignac estimate the intergenerational wealth correlation between two generations for cohorts covering the 20th century. First, they find that the probabilities for the second generation to belong to top wealth groups or to be homeowner increase with the wealth of the parents. Such effects are persistent over the life-cycle. Second, the relative effect of parental wealth is increasing across top wealth groups. Third, the intergenerational correlation in home ownership status is increasing for more recent cohorts. Fourth, the effect of parental wealth on the probability to belong to top wealth groups follows an inverted U-shape over the life-cycle. Fifth, the higher in the wealth distribution, the more important is the role of the receipt of gifts and inheritances and the occupation of the fathers: they fully explain the intergenerational correlation regarding the probability to belong to the top 5%. The education of the household plays an additional role for the top 50%.
Using US Census decennial and American Community Survey data linked to historical IRS Form 1040 records, Akee, Jones, and Simeonova identify household-level eligibility for early years of the Earned Income Tax Credit (EITC). They examine taxpayers beginning in 1979, and link these taxpayers to their children, whose outcomes are examined in later years of census and IRS data. Using a shift-share approach, the researchers examine the impact that differential take-up by geography and demographic group had on the outcomes of children from EITC-eligible families. They also examine the exogenous changes to EITC generosity that rolled out in the late 1980s to mid-1990s, calculating the differential effect of these changes on children from different family sizes and different ages at program change. This research constitutes the first attempt to examine the EITC’s impact on long-term intergenerational income mobility.
Connolly, Haeck, and Laliberté provide evidence on the causal relationship between maternal education and the intergenerational transmission of income. Using a novel linkage between intergenerational income tax data and Census data for individuals born between 1963 and 1985 and their parents, the researchers show that rank mobility has decreased over time, and that this decline was sharpest for children of mothers without a high school diploma. Using variation in compulsory schooling laws, the researchers show that rank mobility increases as the percentage of mothers with a high school diploma increases. They find weaker evidence that mobility increases with the percentage of mothers with a university degree.
Langetieg, Medalia, Meyer, Payne, Plumley, Wu, and Finley provide the first accurate picture of transfer and tax credit receipt in the US along with the distributional consequences of receipt using a groundbreaking set of linked survey and administrative data. The administrative data cover earnings and asset income from IRS tax records and transfer income for a myriad of safety net programs including Social Security, SSI, SNAP, Unemployment Insurance, Veterans’ Benefits, Public Assistance, housing assistance, Medicare, Medicaid, WIC, and energy assistance. The researchers link these data to the Current Population Survey, the source of official poverty and inequality statistics and the Survey of Income and Program Participation, the most comprehensive survey of income sources in the US. Linking the administrative data to the surveys is vital given that a large and rising share of benefits and other income sources is not recorded in the surveys. Using these linked data, the researchers examine the extent to which misreporting of various survey income sources biases estimates of the receipt and distributional consequences of a wide range of transfers and tax credits. A crucial aspect of this work is reconciling differing administrative and survey reports of income sources. This issue is particularly important for earnings where survey reporters may not file tax returns or may misreport their earnings, or firms may not report wages paid.
The distribution of estates (the net value of real and financial property of a deceased person) has commonly served for the estimation of the distribution of wealth among the living via the estate multiplier method, but has never been under extensive scrutiny in and of itself. Theoretically, the application of detailed multipliers can increase or decrease top wealth shares relative to estate shares. This depends on the evolution of mortality rates with respect to age, gender, income and wealth. Alvaredo, Berman, and Morelli highlight that the concentration of estates and the derived concentration of wealth at the top through the multiplier method are actually very close to each other. They investigate why the application of mortality multipliers does not alter significantly the picture when both distributions are compared. The researchers study the general conditions under which the concentration of estates at death provides the same informative content as the concentration of the wealth among the living, and identify that the relationship depends on the covariance between mortality rates and estates amounts. As a result, the researchers provide novel historical series of wealth concentration in country-years where there is no information to apply the multiplier method, but enough data to make well-grounded inferences on the distribution of estates.
Bach, Bartels, and Neef provide a new income inequality series for reunified Germany combining tax data, survey data to national accounts. Estimating distributional national accounts (DINA), they capture 100% of national income and can compute the distribution of pretax and posttax incomes for the entire population. This allows the researchers to answer the following questions: Who has benefited more from economic growth: employees or capital owners? The bottom 50%, the middle class or the top 10%, 1% and 0.1% of earners? Further, this research is the first to apply the DINA methodology to the analysis of regional disparities. 30 years after the German reunification, substantial income differences remain between those living in East and West Germany. In the 1990s, West German investors bought real estate and factories in East Germany, following favourable tax incentives. The researchers investigate to what extent capital income generated in East Germany flowing to West German capital owners can explain structural differences between the income distributions in East and West Germany.
Larrimore, Mortenson, and Splinter present new estimates of the level and persistence of poverty among US households since the Great Recession. They build new annual household data files using US income tax filings between 2007 and 2018. These data, which are constructed for the population of US residents, allow the researchers to track individuals over time and measure how tax policies affect poverty trends. Using an after-tax household income measure, the researchers estimate that over 4 in 10 people spent at least one year in poverty between 2007 and 2018. Those that experienced at least one year of poverty spent an average of one-fourth of the 12-year period in poverty. There is substantial mobility in and out of poverty. For example, 41 percent of those in poverty in 2007 were out of poverty in the following year. However, many of those who are poor spend multiple years in poverty or escape poverty only to fall back into it. Of those who were in poverty in 2007, one-third are in poverty for at least half of the years through 2018. The researchers also document substantial heterogeneity in these trends by age: younger individuals experience higher rates of poverty but less persistence -- older individuals experience lower rates of poverty but more persistence.
Acciari, Alvaredo, and Morelli study the concentration of personal wealth in Italy between 1995 and 2016 using a novel source of data on the full record of inheritance tax files, covering 50-60% of total decedent population. Estimates of the shares accrued to top and bottom wealth groups are derived using the estate multiplier method applied to wealth left at death. The main baseline national series of wealth concentration, derived to be fully consistent with the National Accounts, suggests that richest 1% of Italian adults increased their share of total personal wealth from 20% to 26% approximately from 1995 to 2016. The level of concentration appears in line with other European countries. Data also allow a rich disaggregation of the analysis by demographic and geographic haracteristics. Tax exempt assets as well as unreported wealth hidden in offshore accounts are estimated and distributed back to the population. A range of alternative series of wealth concentration helps to better understand the role of adjustments and imputations in driving the most important findings and allows for better historical comparability of the estimates.
Fessler and Schürz integrate concepts from sociology into the economic analysis of inequality and propose a relational approach focusing on different functions of wealth. They operationalize these functions by empirically analyzing the groups of renters, owners, and capitalists. Employing European and US data, the researchers find that classifying households based on these functions of wealth aligns well with the income and wealth distribution, in ways that vary considerably across countries. The approach allows us to distinguish between wealth as a means of capitalist production, as a substitute for public wealth (precautionary wealth), and as a source of non-cash income (housing wealth used). The researchers propose new measures of inequality directly linked to social realities.
Using Tax Data to Better Capture Top Incomes in Official UK Income Inequality Statistics
The Wealth of Generations, With Special Attention to the Millennials
Distributing Personal Income: Trends Over Time
Presence and Persistence of Poverty in U.S. Tax Data
What Explains the Gender Gap in Wealth? Evidence from Administrative Data
Social Security Wealth, Inequality, and Lifecycle Saving
On the Distribution of Estates and the Distribution of Wealth: Evidence from the Dead
Wealth Transfers and Net Wealth at Death: Evidence from the Italian Inheritance Tax Records 1995–2016
The Accuracy of Tax Imputations: Estimating Tax Liabilities and Credits Using Linked Survey and Administrative Data