Behavioral Responses to Estate Taxation: Evidence from Taiwan (with Tzu-Ting Yang)
Selected for the REStud North America Tour 2025
Abstract: We quantify behavioral responses to estate taxation by exploiting two large reforms in Taiwan. Using comprehensive administrative data and a difference-in-difference design, we show that the response of reported estates to the reforms is quick, persistent, and exhibits an asymmetry. We estimate elasticities of reported estates with respect to the net-of-tax rate of 2.76 (s.e. 0.39) for the tax increase and 1.31 (s.e. 0.16) for the tax cut. The asymmetry arises because liquid items such as financial assets, deposit savings, and charitable exemptions respond significantly more to a tax increase. The quick adjustments in reported estates, combined with a null effect on labor supply behavior among both donors and heirs, suggest the responses are likely driven by tax avoidance. The observed asymmetry can be explained by tax avoidance with sunk costs: taxpayers increase avoidance during a tax increase but are less responsive to a tax cut due to previously incurred avoidance costs. We set up a tax avoidance model and derive sufficient statistics, characterized by our estimated elasticities, to assess the welfare impact of tax reforms. Our analysis shows that using the tax cut elasticity, which is attenuated due to sunk costs, would underestimate the welfare cost and overestimate the net welfare effect of a tax increase by 61%.
Family Businesses, Nepotism, and Productivity (with Hsien-Ming Lien)
Abstract: This paper provides the first systematic evidence on kinship networks within and across businesses by linking comprehensive administrative data on family ties, firm ownership, firm-to-firm transactions, and employer-employee records for the universe of firms in Taiwan. The unique dataset enables us to document the importance of family connections both within and across firms. We document that 40% of firms employ family members and 33% have family owners with relatives owning other firms. Family-linked firms have a 25% likelihood of trading with each other at higher transaction values than non-family partners. These patterns suggest that family ties contribute to shaping business networks and raise questions about their impact on firm performance. Using a stacked event study design, we examine the effects of ownership transfers on firm performance, comparing ownership transfers to family members versus unrelated individuals. While firms transferred within the family experienced little change in sales, firms transferred to unrelated individuals increased sales by over 13% four years later. The improvement can be attributed to the restructuring of employee and supply chain networks, as firms recruit new employees and establish new supplier-buyer relationships.
Selected Work in Progress
