BDBochuan Dai

Bochuan Dai

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Bochuan Dai has a total of 8 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • When Do Narratives Move Prices? Portfolio Allocation and Narrative Diffusion in Financial Markets

    by Yong Ma, Xianxin Duan, Bochuan Dai

    Why do widely discussed market narratives sometimes trigger large asset‐price movements while at other times have little effect? We develop a dynamic equilibrium model in which narratives diffuse through investor communication networks and affect prices through portfolio reallocation. Narratives influence equilibrium prices only when adoption changes perceived asset values sufficiently to alter portfolio demand. The magnitude of their price effects depends on who adopts, is larger when existing narrative valuation is low, and remains bounded by aggregate wealth. Communication networks therefore shape prices indirectly by influencing the distribution of narrative views and investors' adoption decisions. An empirical application to four U.S.‐listed firms associated with the quantum‐computing investment narrative provides complementary reduced‐form evidence. More favorable narrative coverage is associated with higher contemporaneous abnormal returns, while greater narrative intensity is associated with excess volatility and a higher probability of extreme abnormal returns.

  • Open Access

    Industry Risk Decomposition and Mean-Gini Portfolio Choice

    by Bochuan Dai, Yong Ma, Xiang Shi

    Building on the axiomatic subgroup Gini decomposition of Heikkuri and Schief (2026), we reinterpret risky capital as a population partitioned by industry and derive a within-industry benchmark that depends only on industry weights and marginal Gini risk. Portfolio Gini splits into this benchmark and a signed across-industry residual that combines cross-industry hedging with industry-risk heterogeneity. Reweighting this residual changes both components. We study the component-weighted criterion alongside a convex hedging-only benchmark that isolates dependence shrinkage. Both criteria serve as partition-dependent risk regularizers. In controlled simulations that fix the realized mean, partial reweighting yields statistically detectable but economically small improvements that the hedging-only benchmark nearly matches, indicating that dependence shrinkage drives the effect. Sample-only cross-validation selects full weighting, with no improvement in risk-adjusted performance. The controlled gains identify a dependence-regularization effect.

  • Open Access

    From Attention to Verification: Analyst Scrutiny and Corporate AI Washing

    by Bochuan Dai, Shaojie Lai, Zhen Yang Lim, Xiang Shi

    We examine how analyst scrutiny affects corporate AI patenting and AI washing. Using 29,468 firm-year observations for analyst-covered Chinese A-share firms from 2007 to 2024, we find that greater analyst report volume is associated with more AI patent applications and lower AI washing. The negative relationship is stronger when firms receive site visits from the brokerages that cover them and when AI is raised in the questions asked at those visits. The positive association with AI patenting and the negative association with AI washing are stronger when analysts are more experienced, brokerages have greater resources, and firms employ more technical and highly educated workers. By linking analyst reports with brokerage visits and AI-specific inquiry, we connect the intensity of analyst scrutiny to observable opportunities to evaluate corporate AI claims.

  • Open Access

    AI Adoption and Firm-Level Wage Structures: Evidence from China

    by Bochuan Dai, Zhen Yang Lim, Xiang Shi, Lu Wang

    This paper examines how artificial intelligence (AI) adoption affects firm-level wage outcomes in Chinese listed firms. Using 30,560 firm-year observations from 2011 to 2023, we measure AI adoption through textual analysis of annual reports. We find that AI adoption is positively associated with both employee wages and management wages. The results remain robust across instrumental variable estimation, propensity score matching, a policy-based difference-indifferences design, and alternative AI measures. We further show that workforce restructuring is an important channel: AI adoption increases the share of highly educated employees while reducing reliance on lower-educated workers. In addition, the wage effect is stronger in non-state-owned enterprises, and AI adoption weakens pay-for-performance sensitivity.