Research

1. MS

   

   Data-Driven Vertical Integration: Evidence from Large-Scale Job Postings

   Zexuan Liu, Zhiyuan Chen, Fuxin Zhai, and 1 more author

   2026

   Revise and Resubmit at Management Science

   Abs Bib

   We investigate how data technology influences vertical integration in China’s production network, using a dataset of 99.32 million job postings and firm registration records for 52.6 million companies (2017-2019). By leveraging machine learning algorithms, we precisely map job titles to their corresponding data processing requirements. Using OLS and shift-share instruments, we find: (1) increased data technology adoption in downstream sectors boosts vertical integration, particularly backward integration; (2) contract litigation predicts integration but doesn’t fully explain the effect of data technology; (3) producing firms target data-intensive industries, with most integration happening in software and information services sectors.

   @unpublished{liu2026vertical,
     title = {Data-Driven Vertical Integration: Evidence from Large-Scale Job Postings},
     author = {Liu, Zexuan and Chen, Zhiyuan and Zhai, Fuxin and Chih, Yaoyu},
     note = {Revise and Resubmit at Management Science},
     year = {2026},
   }

2. SMJ

   

   Learning without Experience: Data Technology and Firm Scope

   Zexuan Liu, Liang Chen, and Zhiyuan Chen

   2026

   Revise and Resubmit at Strategic Management Journal

   Abs Bib

   Organizational learning theory and evidence suggests that firms traditionally expand into new markets cautiously, accumulating experience through trial-and-error before venturing into unfamiliar markets. We investigate whether modern data technology can alter this paradigm by enabling firms to "learn without experience" when extending their geographic scope. Adopting a question-driven, abductive approach, we examine a novel panel of Chinese firms from 2017-2019, a period of rapid diffusion of data analytics tools. Using nearly 100 million job postings to identify data-technology adoption and a shift-share (Bartik) instrumental variable for causal inference, we find that firms adopting data technology enter on average 2.3 more new markets and expand about 600 km farther from their home region than non-adopters. Crucially, firms equipped with data technology no longer avoid entirely new or distant markets, deviating from the traditional stepwise expansion path. To interpret this strategic shift, we develop a formal model that considers two potential mechanisms - mitigating market uncertainty versus improving operational efficiency. Additional analyses provide strong evidence that points to uncertainty mitigation as the primary driver. By depicting data technology as a substitute for experiential learning in coping with external uncertainty, our study reveals how data-driven insights help broaden firm scope - a fundamental question for strategy scholars in the digital age.

   @unpublished{liu2026scope,
     title = {Learning without Experience: Data Technology and Firm Scope},
     author = {Liu, Zexuan and Chen, Liang and Chen, Zhiyuan},
     note = {Revise and Resubmit at Strategic Management Journal},
     year = {2026},
   }

2026

1. 经济研究

   

   中国专利质量的测度、变化机制与经济增长效应

   Zhiyuan Chen, Yiran Zhang, Jie Zhang, and 1 more author

   经济研究, 2026

   Abs Bib

   如何构建有效测算中国专利质量的方法体系，既是该领域的前沿问题之一，也是事关中国如何全面落实创新驱动发展战略的基础性工作。本文利用1985—2021年期间中国发明专利文本信息，运用自然语言处理方法，在计算符合中国专利间文本相似度这一核心指标的基础上，探讨了适用于中国专利数据特征的专利质量测算方法。各项可靠性检验均验证了本文设计和测算的专利质量指标变量相较于传统专利质量指标的优势。基于这一专利质量指标，本文从专利申请人类型、地域分布和技术分类等角度梳理和归纳了中国1991—2016年期间中国专利质量水平的变化趋势，揭示了中国专利质量全面提升的重要特征事实。在此基础上，利用包含专利质量和数量变化的GDP增长分解框架，探究专利质量上升对中国经济增长的贡献度，发现样本期间内专利质量对经济增长的平均贡献度约为3%，且其对经济增长的作用相较专利数量呈逐步增强态势。本研究既可为有效评估各类微观主体的专利真实市场价值和自主创新能力提供基础性的识别工具，更可为中国一系列专利制度改革政策及各类科技创新激励政策的绩效评估提供参考依据。

   @article{liu2026patent_quality,
     title = {中国专利质量的测度、变化机制与经济增长效应},
     author = {Chen, Zhiyuan and Zhang, Yiran and Zhang, Jie and Liu, Zexuan},
     journal = {经济研究},
     year = {2026},
     number = {03},
   }

2024

1. JHC

   

   Computerization and the Decline of Unincorporated Self-Employment

   Zexuan Liu^*, Sun Lee, and Yao-yu Chih

   Journal of Human Capital, 2024

   Abs Bib HTML

   We investigate the effect of the computer adoption rate, referred to as computerization, on the proportion of unincorporated self-employed individuals within the US labor markets. The conceptual framework suggests that computerization may either augment or diminish this share. Employing a Bartik instrument approach, we disentangle the causal effect of computerization on unincorporated self-employment from 1990 to 2010. Our empirical findings indicate that a 1% increase in computerization corresponds to a 0.79% reduction in the unincorporated self-employment share.

   @article{liu2024computerization,
     title = {Computerization and the Decline of Unincorporated Self-Employment},
     author = {Liu, Zexuan and Lee, Sun and Chih, Yao-yu},
     journal = {Journal of Human Capital},
     year = {2024},
   }

2023

1. EmpEcon

   

   The Puzzle of Changes in Employment and Wages in Routine Task-Intensive Occupations

   Pallab Ghosh and Zexuan Liu^*

   Empirical Economics, 2023

   Abs Bib HTML

   Autor and Dorn (Am Econ Rev 103(5): 1553–1597, 2013) provide an explanation of the polarization of US employment and wages for the period 1980–2005. Using the 1980 Census and 2005 American Community Survey data, this study replicates the estimation results of Autor and Dorn (2013) for employment polarization in all major occupation groups and qualitatively matches the wage polarization results. Also, we investigate the puzzle of why employment and wages changed in opposite directions only in clerical and administrative support occupations among three high routine task-intensive occupations.

   @article{liu2023puzzle,
     title = {The Puzzle of Changes in Employment and Wages in Routine Task-Intensive Occupations},
     author = {Ghosh, Pallab and Liu, Zexuan},
     journal = {Empirical Economics},
     year = {2023},
   }

2022

1. JATM

   

   Incumbents’ Pricing and Non-pricing Responses to Entry in Vertically Differentiated Markets

   Christoph Bauner and Zexuan Liu^*

   Journal of Air Transport Management, 2022

   Abs Bib HTML

   We investigate incumbents’ pricing and nonpricing responses to product entry in market with vertical product differentiation. Using data from international airline markets, we observe price reductions by connecting incumbents in response to high-quality (nonstop) entry. We also find evidence of product repositioning (increases in total itinerary distance) by connecting incumbents. However, we do not observe significant policy changes by incumbents offering nonstop flights. Additional evidence suggests that the reason for the response by connecting incumbents is the opportunity cost of connecting at desirable airports such as airline hubs.

   @article{liu2022incumbents,
     title = {Incumbents' Pricing and Non-pricing Responses to Entry in Vertically Differentiated Markets},
     author = {Bauner, Christoph and Liu, Zexuan},
     journal = {Journal of Air Transport Management},
     year = {2022},
   }

2021

1. EI

   

   Heterogeneous Price Effects and Increased Price Dispersion from Quantity-based Congestion Management

   Sun Lee, Zexuan Liu^*, and Haojun Yu

   Economic Inquiry, 2021

   Abs Bib HTML

   We examine the effects of quantity-based approaches to airport congestion management such as slot restrictions on price dispersion and on-time performance. From the evidence at Newark Liberty International Airport, we find that slot restrictions increase price dispersion and improve on-time performance. Price dispersion originates from heterogeneous price effects, as ticket prices for high-value passengers increase more than that for low-value passengers.

   @article{liu2021heterogeneous,
     title = {Heterogeneous Price Effects and Increased Price Dispersion from Quantity-based Congestion Management},
     author = {Lee, Sun and Liu, Zexuan and Yu, Haojun},
     journal = {Economic Inquiry},
     year = {2021},
   }

2020

1. WD

   

   Effects of Motherhood Timing, Breast-milk Substitutes and Education on the Duration of Breastfeeding: Evidence from Egypt

   Firat Demir, Pallab Ghosh, and Zexuan Liu^*

   World Development, 2020

   Abs Bib HTML

   Breastfeeding has significant health and human capital effects on both mothers and infants. However, breastfeeding rates vary significantly within and across countries as societal, political, economic and cultural factors along with individual choices shape the breastfeeding practices. Using data from the Egyptian Demographic and Health Surveys, this study examines the effects of first motherhood timing, availability of breastmilk substitutes, and mothers’ education levels on breastfeeding duration in a major developing country, Egypt. The empirical analysis, which corrects for the estimation errors that plagued previous research, shows that delaying the first motherhood timing and increasing the availability of infant formulas have statistically significant negative effects on breastfeeding duration. Furthermore, breastfeeding duration is found to be decreasing in mothers’ education levels.

   @article{liu2020breastfeeding,
     title = {Effects of Motherhood Timing, Breast-milk Substitutes and Education on the Duration of Breastfeeding: Evidence from Egypt},
     author = {Demir, Firat and Ghosh, Pallab and Liu, Zexuan},
     journal = {World Development},
     year = {2020},
   }

2019

1. EL

   

   Ownership and Pricing Response to Entry

   Zexuan Liu^*

   Economics Letters, 2019

   Abs Bib HTML

   I explore state-owned and privately-owned incumbents’ differential pricing responses to low cost carrier (LCC) entry in the international airline market. I find that privately-owned incumbents substantively cut prices in response to LCC entry, whereas state-owned incumbents do not. The differential responses of state-owned vs. privately-owned is more considerable on the slot-restricted routes, where incumbents have monopolistic power. In addition, Chinese airlines and Gulf carriers who are aggressively expanding behave differently from other state-owned carriers.

   @article{liu2019ownership,
     title = {Ownership and Pricing Response to Entry},
     author = {Liu, Zexuan},
     journal = {Economics Letters},
     year = {2019},
   }

2. SEJ

   

   Do State Minimum Wages Affect the Incarceration Rate?

   Pallab Ghosh, Gary Hoover, and Zexuan Liu

   Southern Economic Journal, 2019

   Abs Bib HTML

   Because of historically unprecedented increases in the prison population since the late 1970s, the United States has the highest incarceration rate in the world. Using Becker’s (1968) framework on crime, this study investigates the causal relationship between incarceration rates and state minimum wages. Our identification strategy consists of a state-level fixed effects model and Autor and Dorn’s (2013) two-stage least-squares (2SLS) approach. Using the historical local industry structure, we predict the change in employment shares of manual task-intensive occupations and use those as an instrument for state minimum wages. The fixed effects and 2SLS estimates suggest that a one-dollar increase in state minimum wage leads to approximately 12–25 fewer incarcerations per 100,000 state residents. Estimates of the heterogeneous impacts by race and gender indicate that the aggregate impact of state minimum wages is entirely driven by men.

   @article{liu2019minimum,
     title = {Do State Minimum Wages Affect the Incarceration Rate?},
     author = {Ghosh, Pallab and Hoover, Gary and Liu, Zexuan},
     journal = {Southern Economic Journal},
     year = {2019},
   }

3. AEL

   

   Coauthorship and the Gender Gap in Top Economics Journal Publications

   Pallab Ghosh and Zexuan Liu^*

   Applied Economics Letters, 2019

   Abs Bib HTML

   Coauthorship has become the new norm in academics since the early 1970s because it enhances productivity. However, the gender gap in economics journal publication continues to persist, which explains why there are fewer women than men in the economics profession. This study investigates the role of coauthorship in the gender gap in top economics journals for untenured faculty members in the US. We construct a unique dataset from the CVs of academic economists from the top 96 PhD-granting economics departments in the US. Our results suggest that, compared to men, women are matched with less productive unique coauthors because women begin their academic careers in lower-ranked economics departments than men, which is associated with institutional gender bias. This poor-quality matching can explain approximately 0.60 fewer publications in the top 20 economics journals during the untenured period.

   @article{liu2019coauthorship,
     title = {Coauthorship and the Gender Gap in Top Economics Journal Publications},
     author = {Ghosh, Pallab and Liu, Zexuan},
     journal = {Applied Economics Letters},
     year = {2019},
   }