Works in Progress

"Machine Learning for Treatment Effect Heterogeneity: Recovering Partial Effects". Joint with Elad Guttman, Itay Saporta-Eksten and Analia Schlosser.

Working Papers

"Conditional Triple Difference-in-Differences". arXiv link

Abstract:  Triple difference-in-differences (TDID) designs are widely used in empirical research to estimate causal effects. In practice, most implementations rely on a specification with controls. However, we show that such approaches introduce bias due to differences in covariate distributions across groups. To address this issue, we propose a re-weighted estimator that correctly identifies a causal estimand of interest by aligning covariate distributions across groups. For estimation we develop a double-robust approach. A R package is provided for general use.

"Correcting invalid regression discontinuity designs with multiple time period data". Joint with Daniel Nevo. arXiv link

Abstract:  A common approach to Regression Discontinuity (RD) designs relies on a continuity assumption of the mean potential outcomes at the cutoff defining the RD design.  In practice, this assumption is often implausible when changes other than the intervention of interest occur at the cutoff (e.g., other policies are implemented at the same cutoff). When the continuity assumption is implausible, researchers often retreat to ad-hoc analyses that are not supported by any theory and yield results with unclear causal interpretation. These analyses seek to exploit additional data where either all units are  treated or all units are untreated (regardless of their running variable value). For example, when data from multiple time periods are available. We first derive the bias of RD designs when the continuity assumption does not hold. We then present a theoretical foundation for analyses using multiple time periods by the means of a general identification framework incorporating data from additional time periods to overcome the bias. We discuss this framework under various RD designs, and also extend our work to carry-over effects and time-varying running variables. We develop local linear regression estimators, bias correction procedures, and standard errors that are robust to bias-correction for the multiple period setup. The approach is illustrated using an application that studied the effect of new fiscal laws on debt of Italian municipalities. 

Published Papers

Danieli, O., Gilat, S., & Leventer, D. (2024). "Top Income Inequality in Israel". Israel Economic Quarterly, 22(1), 157-244.  Link

Abstract:  This paper is the first to estimate top income inequality in Israel using administrative micro-data. Combining employee and self-employed tax-records from 2008-2018, we estimate that on average the top 1% earned 14.7% of the total income during this period, a relatively high estimate compared to other OECD countries. Decomposing by income type we find that for the top 1-0.1%, income originates mainly from labor and business income, and for the top 0.1%  the main income source is capital income, and primarily dividends. During the last decade, and mostly after 2015, top income shares slightly decreased, mostly due to a decline in labor income inequality. Classifying economic industries of top earners, we find that the most common industries in the top 1-0.05% are Medical practices, High-tech and Legal services. For the top 0.05\% these are Management Consultancy, Wholesale trade, High-tech and Real-estate industries. Lastly, we estimate intragenerational mobility rates between 2008 and 2018. We document a 5% probability to enter the top decile from the bottom nine deciles, and a 77% probability to stay in the top decile if started in the top 1%.

Masters Thesis

Abstract: Studies have extensively documented cross-sectional racial gaps between Arabs and Jews in Israel, but less attention has been given to income gaps from an intergenerational perspective. Using paired child-parent administrative data on the universe of Israeli children from birth cohorts 1975-1988, we  document four empirical findings. First, using rank-rank correlations we calculate a long-term racial income rank gap of 12 income ranks. Considering birth-cohorts separately reveals that the long-term racial income inequality is decreasing, since Jews are at a constant long-term mean racial income rank, but the Arab's long-term mean racial income rank is on an increasing trajectory. Second, to better understand the long-term racial gaps we delve into the foundations of the intergenerational racial income gap and nd that it stems mostly from the gap between Arab and Jewish women at the lower tail of parent income. Third, we consider how a larger set of child outcomes account for the intergenerational racial income gap and nd using re-weighting exercises that labor supply and education account for a significant share of the gap between Jewish and Arab women, but number of children in the children's generation accounts for a negligible share. Fourth, we provide evidence that childhood level factors other than parent income, e.g. parent education, account for a substantial amount of the racial income and education gaps at the lower tail of parent income for both genders. But, we find that for women similar childhood environments do not account for the labor supply disparities.