Research
Working Paper
The Race between Aging and Automation: Age-biased Technology and Age Pay Gap (JMP)
Abstract
Motivated by the concurrence of workforce aging and automation, this paper examines how technological changes reshape the wage gap between older and younger workers in the U.S. labor market. I document that the U.S. age pay gap expanded continuously until the mid-2010s before experiencing a sharp compression. To explain this age pay gap reversal, I develop a task-assignment model featuring age-heterogeneous workers, where older incumbents hold a comparative advantage in non-routine cognitive tasks. The model distinguishes between Positive Age-Biased Technologies (PABT, e.g., software, robotics, and machines), which automate routine tasks and complement experience, and Negative Age-Biased Technologies (NABT, e.g., artificial intelligence), which directly automate non-routine cognitive tasks. To quantify these mechanisms, I construct a novel occupational technology-task exposure measure by applying a fine-tuned Sentence-BERT model to match patent abstracts with job task descriptions. Micro-level estimates reveal a dual impact: PABT generates a massive wage premium for older workers, whereas NABT imposes severe penalties on their wage growth trajectories. Structural estimation confirms that age pay gap is jointly driven by technological and demographic changes. It further documents that an aging demographic structure amplifies intergenerational wage inequality under PABT but accelerates the gap's compression under the NABT change. These findings provide new insights for flexible retirement design and technology regulation in the context of the ongoing race between demographics and technology.Fertility Effects of Parental Early Retirement
Abstract
This study exploits the eligibility age for early retirement in a regression discontinuity design to investigate the impacts of parental early retirement on the fertility of their adult children. Utilizing data from the Health and Retirement Study, it reveals a rise in the likelihood of daughters having childbirth upon their parents' early retirement. Extensive margins mainly drive this increase, as the effects are especially strong among those expecting their first child. These effects are short-term, lasting up to two years after parental retirement. By exploring time and money transfers between parents and their children, the rise in childbirth can be attributed to increased time transfers from retired parents to their adult children in informal childcare, rather than monetary transfers. Moreover, the increase in childbirth after early parental retirement is more pronounced in areas with higher formal childcare costs.The Health Effects of Paid Family Leave joint work with X. Wang
Abstract
This paper examines the effects of New Jersey's Paid Family Leave (NJ-PFL) program on maternal and infant health and whether these effects vary by race and access to job protection. Using restricted-use natality data from the National Vital Statistics System for 2006--2011, we estimate event-study and difference-in-differences designs comparing New Jersey with neighboring states before and after the program's implementation in 2009. NJ-PFL increased prenatal care utilization by 0.58 visits and reduced preterm delivery by 0.54 percentage points, or 4.2\% relative to the pre-policy mean. Point estimates suggest larger improvements in prenatal care and preterm delivery among non-Hispanic Black mothers, as well as larger reductions in pregnancy-associated hypertension. Estimated benefits are also greater in areas with higher potential job-protection coverage, although the racial and geographic differences are imprecisely estimated under conservative inference. These findings highlight the heterogeneous effects of PFL on maternal and infant health and the potential importance of combining wage replacement with job protection to promote equitable access to leave.
Selected Work in Progress (by Topic)
1. Institutions, Labor Markets, and Family Decisions
Retirement Slowdown and Fertility Decline: Evidence from the U.S. (draft coming soon)
Abstract
Rapid demographic aging is concurrent with declining fertility rates in many countries, but the causality has not been well understood so far. Leveraging variations in historical age structure as Bartick-IV for local retirement flow, and birth outcomes from vital records over period 1989-2023, the study finds no evidence that the less newborns are in commuting zones with fewer older workers retiring. However, it finds less mothers with college degrees, more older mothers. This suggests that the retirement slowdown has delayed, or even redistribute younger generation's birth.When Balance Creates Imbalance: Unexpected Effects of Gender Quota in Hiring
Highlight
I study the 1:1 gender quota in China's civil service recruitment in a staggered difference-in-differences design and find it crowds out female candidates from labor participation in public sectors, and therefore delaying marriage and fertility, especially in counties with higher gender ratio.Fertility Responses to Paid Family Leave, with Xueshi Wang
Highlight
Using the U.S. vital records and exploiting a discountinuity design, we find paid family leave leads to a spike in birth number, and C-section deliveries within a 10-month window after the introduction of this policy,particularly among underpresented groups.Household Effects of a Shorter Workweek: Marriage, Fertility and Parenting, with Hailun Wei, and Tianli Yang
Highlight
We investigate the impacts of a shorter workweek (from 6 days to 5 days) quasi-experiment, and document job amenities with one day off in 3 sets of differences in difference designs that: (1) in a short term of 1-year window, it increases likehood of getting married and having a newborn per young (i.e., 20-30) female in the late 1990s of China; (2) in a middle term, children (aged 12-14) exposed to this reform increases the probability of compltementing high schools, and teenage (aged 15-17) exposure to this reform increases the probability of completing college; (3) in a longer term, young kids (aged 0-11) exposed to this reform has a higher likelihood to get married and birth by their 30s.Dual-Track Family Crashes after Anti-Corruption Campaigns
Highlight
The "dual-track in the family", whereby one spouse employed in public sector enables the other to pursue private entrepreneurship, declines after a national wide anticorruption campaign. This offers a new understanding of this dual track in the family previously thought to underpin private-sector dynamism by leveraging state-sector stability: it's also the micro-level foundation of privatizing public resources.Land into Marriage: Effects of Land Decollectivization Reform in China
Highlight
Land decollective reform (Household Responsibility System since 1978, HRS hereafter) provides farmers with autonomy over the means of production, and brings in agriculture expansion. However, rare evidence relates to how it reshapes marriage and family upon this property right reform. With Census data, I find in counties adopted HRS, the marriage rate among females between 16 and 50 increases sharply, mostly driven by those teenage marriage (aged 16-20) and late-marriage (30 and over): marriage serves as a means of gaining access to and using productive resources.2. Technology(Patent Documents)and Textual Analysis
The Divergence of Goods and Knowledge Flows: Evidence from Trade War, with Jiawen Liu
Highlight
Exploiting fine-tuned Language Models to classify trade and patent records, we observe the declining commodity flows but increasing patent flows related to targeted goods between countries involved in a unilateral trade war. In contrast, only the domestic patent numbers related to regulated goods increase in a bilateral trade war where decoupling effects outweigh goods-knowledge substitution effects.
Publication
Liu, J., and X. Liu. 2026. Spillover of Female Labor Protection: Evidence from China. Journal of Labor Research. 47(2).
Liu, Y., X. Liu, Y. Zhang, and K. Wang. 2025, The Tax Compliance Effect of Banking Competition. Economic Science. (in Chinese) 5, 98-122.
Tsai, P-H, Y,Liu, X. Liu. 2021. Collusion, Political Connection, and Tax Avoidance in China. Kyklos. 74: 417–441.
Liu, X., Y. Liu. 2021. Land Lease Revenue Windfalls and Local Tax Policy in China. International Tax and Public Finance. 28: 405–433.