Research Papers

We evaluated inequality of opportunity across 31 EU countries by applying Bayesian model averaging and tree-based machine learning, focusing on how uncertainty in the choice of circumstance variables and functional forms affects estimates.

Standard “kitchen sink” models often overstate the role of circumstances in shaping income outcomes. By systematically addressing model uncertainty, we sought to produce more accurate and robust measures that reflect how social class, parental education, and household conditions shape opportunity.

Our findings show that ignoring model uncertainty can mislead policy by exaggerating the extent of inequality of opportunity. Recognizing the roles of class, education, and migrant background informs better strategies for reducing unjust disparities across societies.

We examine intergenerational educational and occupational mobility in China and Russia during their transitions from centrally planned to market economies. We focus on the extent new generations can escape institutional and structural constraints shaping their parents’ outcomes. We estimate overall and steady-state mobility and decompose it to reveal mechanisms driving intergenerational change.

Analysis of intergenerational mobility in China and Russia must account for the fact that observed mobility arises within an evolving dynamic system. We offer the first attempt to measure intergenerational mobility in these transition economies, isolating the transmission of educational and occupational status between parents and children from the direct effects of exogenous structural change.

We offer a novel perspective on intergenerational mobility in transition economies. Our measures address limits of traditional approaches that overstate mobility in contexts like China and Russia. By isolating transmission of social status from structural changes, this method allows more accurate assessments and enhances comparability of mobility across countries.

We conducted the first systematic empirical analysis of income inequality across Chinese provinces using CHFS data and a semiparametric model to estimate Gini indices from 2012 to 2018, identifying human capital differences as key drivers.

We aimed to understand intraprovincial inequality, which accounts for 90% of national income variation and contributes to economic disparity, slower growth, and potential instability.

Our findings suggest that government policies promoting educational development in disadvantaged provinces and reducing barriers to labor mobility can facilitate labor market integration, narrow income gaps, and support sustainable and balanced regional economic development in China.

I estimate income inequality in China from 2012–2018 using new methods that adjust for sparse top-income data in the China Household Finance Survey. I find Gini index estimates of 0.616 (2012), 0.604 (2014), 0.581 (2016), and 0.590 (2018). I also document rural-urban, regional, and provincial inequality patterns in Shanghai, Guangdong, Liaoning, Henan, and Gansu.

Academic and policy discussions are complicated by a lack of consensus on income inequality in China. The existing discrepancies result from the data and methodological issues. In response, I provide estimates of income inequality levels in China using a dataset designed to oversample rich households, and by developing new robust measurement procedures applicable to the datasets with extreme high-income values.

Most researchers share the view that existing high levels of income inequality pose significant warning signs for future economic progress and remain an important challenge for Chinese society and its policymakers. New empirical evidence on the levels and trends of income inequality in China can help inform the design of effective policy interventions.

We developed a Structural Mixed Autoregression (SMAR) model to analyze how macroeconomic aggregates and income distribution interact dynamically. By incorporating functional variables, we measured the effects of monetary policy and income distribution shocks on macroeconomic outcomes. Our model identifies how shifts in income distribution impact aggregate output and how macroeconomic shocks influence inequality.

Existing studies often treat income distribution and macroeconomic aggregates separately. We sought to understand their joint dynamics to uncover whether redistributive policies could enhance growth while reducing inequality. Our approach addresses gaps in traditional models by capturing the full distributional effects rather than focusing solely on mean or median outcomes.

Redistributive policies can boost growth without harming income equality, challenging the growth-equity trade-off. Monetary policy shocks have mixed effects: contractionary policies may reduce inequality but often worsen it by lowering aggregate income. These findings guide policies balancing macroeconomic stability and income equality.

We introduced the General Gender Gap (GGG), a new framework for measuring gender disparities across multiple outcome dimensions, including wages, health, and leisure. This approach improves upon traditional univariate measures by considering substitution and complementarity between different aspects of well-being, providing a more comprehensive view of gender inequality.

Existing gender gap measures primarily focus on singular variables like income, failing to capture the multidimensional nature of well-being. By incorporating multiple outcomes, we aimed to develop a more accurate and policy-relevant measure of gender disparities that reflects the real trade-offs individuals face in economic and social life.

Our findings show that gender disparities in the U.S. differ significantly when examined through the GGG framework, particularly in the trade-off between income and leisure. This new measure offers critical insights for policymakers, enabling more effective interventions to address gender inequality in well-being beyond traditional income-based assessments.

We developed a nonparametric model to study intergenerational income mobility, allowing for nonlinearities and interactions among factors like race, education, and parental age. Using PSID data, we analyzed how these dynamics affect children’s economic outcomes.

Standard models often overlook complex interactions in mobility. Our approach captures how race and parental education shape opportunities, offering a more nuanced understanding of barriers to economic mobility.

Our findings reveal how interactions between parental income, race, and education create mobility bottlenecks, providing insights for policies aimed at reducing economic inequality and promoting upward mobility.

We developed new measures of intergenerational persistence using “memory curves” to track how socioeconomic origins influence mobility across generations. Using U.S. data, we analyzed how initial conditions shape transitions between occupational classes and aggregated insights to describe mobility trends in the population.

Understanding intergenerational mobility requires tools that capture the long-term influence of socioeconomic origins. By modeling these dynamics as “memory,” we aim to better explain how class persistence varies and uncover patterns missed by traditional metrics.

Our findings show that class origins often fade within three generations, but mobility varies significantly by occupational background. These insights help inform policies aimed at addressing inequality and promoting equal opportunities across diverse socioeconomic contexts.

We analyzed a generalized mobility model using administrative data from Denmark to examine the roles of family selection and locational characteristics in generating spatial heterogeneity in intergenerational mobility across nearly 300 larger and 2,000 smaller neighborhoods.

We aimed to understand why intergenerational mobility varies across neighborhoods, an essential issue for shaping economic and social policies.

Understanding the mechanisms behind spatial differences in mobility helps policymakers target interventions to improve equality of opportunity across different neighborhoods.

We combined a special class of deep neural networks with directed acyclic graphs to improve our ability to draw inferences about entire systems of causal relationships.

Social science theories often postulate causal relationships among a set of variables or events, but when it comes to empirically evaluating them, researchers typically resort to methods that impose unrealistic assumptions on the data, leading to incorrect inferences. Our approach can model these relationships without invoking such restrictive assumptions.

Our approach allows researchers to more accurately learn causal relationships from data. It therefore has a very wide scope for application, ranging from studies of inequality and social mobility to health, education, and psychology.