In the US, Asians are commonly viewed as the “model minority” in business academia. Some inspiring initiatives intended to help ethnic minorities to attain tenure and deanship exclude Asians from participating, perhaps because Asians are assumed to be already successful. I challenge this assumption by revealing a “Bamboo Ceiling” in tenure, full professorship, and deanship in US business schools. I analyze a 10-year panel of tenure-track professors and deans at top-50 US business schools. Although Asians appear well represented at first glance, a stark contrast emerges once I distinguish between East Asians (e.g., ethnic Chinese) and South Asians (e.g., ethnic Indians): Among all ethnicities, East Asian faculty are proportionally the least likely to be tenured professors, full professors, and deans, whereas South Asian faculty are the most likely. Moreover, East Asians tend to be employed by lower-ranked schools. To understand these puzzling patterns, I construct large-scale datasets to test potential contributing factors, including (a) faculty recruitment bar, (b) research productivity, (c) research impact, (d) teaching evaluations, (e) invited seminar talks, (f) social media activities, and (g) social media mentions. As one of the largest endeavors to examine ethnic disparities in academia, this research extends the diversity, equity, and inclusion literature and the “leaky pipeline” literature by uncovering East Asian faculty’s neglected challenges in US business schools.

Cross-sectional anomalies and time-series market returns are jointly determined in equilibrium, suggesting a necessity to unify the two central strands of literature for more general understanding of asset pricing. Examining 44 non-U.S. countries, we find that representative cross-sectional anomalies are mostly insignificant at the country level, but become significant once aggregated to the regional level. After regional aggregation, supposedly “market-neutral” long-minus-short anomaly returns predict developed-market returns, while “market-exposed” long-or-short anomaly returns predict emerging-market returns. In contrast, characteristics—foundational to cross-sectional predictability—have no power in time-series predictability. We address data-mining in testing the international anomaly-market linkage by assuming anomalies and their market-return predictability are both data-mined domestically. We rationalize international differences of such linkage by decomposing it into novel measures of foundational market efficiency concepts, including inter-temporal, systematic importance of mispricing, relative importance of efficient-v.s.-inefficient pricing, and asymmetric mispricing correction speed. The first factor and the latter two factors shape the linkage differently across markets of different maturity.

In this talk, I will briefly present two studies that link AI phenomena to conventional theories and demonstrate ways to make theoretical contributions in an area that appears mainly phenomenon-driven so far. The first study theorizes about AI in platform markets. We analyze a model demonstrating how an entrant with a superior algorithm technology may outcompete an incumbent possessing a user base advantage, a strategy we term as “algorithm envelopment.” Our study extends the recent debate on “data network effects” and shows that teasing apart network effects, data-driven learning, and algorithm technology can yield new insights into platform competition, platforms’ corporate-level strategies, and antitrust policymaking. In the second study, we examine the impact of Artificial Intelligence-Generated Content (AIGC) technologies on knowledge search. Our analysis of data from StackOverflow, a popular platform for crowdsourcing coding knowledge, shows that developers are more likely to obtain potential solutions after the rise of ChatGPT. Yet the likelihood of obtaining satisfactory solutions remains unchanged and instead it takes longer to locate satisfactory solutions. A closer look unveils substantial heterogeneity, possibly associated with varying capabilities of solution evaluation, an understudied factor in the search literature.