This paper develops an autonomous framework for systematic factor investing via agentic AI. Rather than relying on sequential manual prompts, our approach operationalizes the model as a self-directed engine that endogenously formulates interpretable trading signals. To mitigate data snooping biases, this closed-loop system imposes strict empirical discipline through out-of-sample validation and economic rationale requirements. Applying this methodology to the U.S. equity market, we document that long-short portfolios formed on the simple linear combination of signals deliver an annualized Sharpe ratio of 2.75 and a return of 54.81%. Finally, our empirics demonstrate that self-evolving AI offers a scalable and interpretable paradigm.

This study investigates the role of contextual cues in memory-based decision-making within high-stakes trading environments. Using trade records from a large Chinese brokerage firm, we provide evidence that both extreme events (COVID-19 quarantines) and everyday contexts (geographic locations) trigger the recall of previously traded stocks, increasing the likelihood of subsequent orders for those stocks. The observed patterns align more closely with similarity-based recall than with alternative channels. Welfare analysis reveals that these memory-induced trades lead to substantial losses for the representative investor's portfolio. We also find evidence at the market level: when the geographical distribution of quarantine risks is recalled, the probability of recalling the cross-sectional stock return-volume distribution from the same day increases by 1.6 percentage points. This study provides evidence from a real-world setting for memory-based theories, particularly similarity-based recall, and highlights a novel channel through which contextual cues affect financial markets.

Can fully agentic AI nowcast stock returns? We deploy a state-of-the-art Large Language Model to evaluate the attractiveness of each Russell 1000 stock each trading day, starting in April 2025 when AI web interfaces enabled real-time search. Our data contribution is unique along three dimensions. First, the nowcasting framework is completely out-of-sample and free of look-ahead bias by construction: predictions are collected at the current edge of time, ensuring the AI has no knowledge of future outcomes. Second, this temporal design is irreproducible once the information environment passes. Third, our framework is fully agentic: we do not feed the model curated news or disclosures; it autonomously searches the web, filters sources, and synthesises information into quantitative predictions. We find that AI possesses genuine stock-selection ability, but that its predictive power is concentrated in identifying future winners. A daily value-weighted portfolio of the 20 highestranked stocks earns a Fama-French five-factor plus momentum alpha of 19.4 basis points and an annualised Sharpe ratio of 2.68 over April 2025–March 2026. The same portfolio accumulates roughly 49.0% cumulative return, versus 21.2% for the Russell 1000 benchmark. The strategy is economically implementable: the average bid-ask spread of the daily Top-20 portfolio is 1.79 basis points, less than 10% of gross daily alpha. However, the signal remains asymmetric. Bottom-ranked portfolios generally exhibit alphas close to zero, while the strongest predictive content sits in the extreme top ranks. Delayed-entry tests further show that predictability does not vanish after a single day; rather, the signal remains positive over a broad window of subsequent entry dates, consistent with slow information diffusion rather than a fleeting overnight anomaly.

This study investigates the role of contextual cues in memory-based decision-making within high-stakestrading environments. Using trade records from a large Chinese brokerage firm and a novel dataset on COVID-19 quarantines, we find that quarantine periods trigger the recall of previously traded stocks, increasing the likelihood of subsequent orders for those stocks. The observed patterns align more closely with similarity-based recall than with alternative channels. Welfare analysis reveals that these memory-induced trades lead to an annualized loss of approximately 70 percentage points for the representative investor’s portfolio. We also find evidence at the market level: when the geographical distribution of quarantine risks is recalled, the probability of recalling the cross-sectional stock return-volume distribution from the same day increases by 1.6 percentage points. This study provides causal evidence from a real-world setting for memory-based theories, particularly similarity-based recall, and highlights a novel channel through which COVID-19 policies affect financial markets.

This paper investigates how foreign institutional investors, specifically Qualified Foreign Institutional Investors (QFIIs), influence the investment strategies of Chinese mutual fund management companies (FMCs) in which they hold shares. By analysing panel data from 1,766 mutual funds managed by 44 foreign-invested FMCs in China between 2005 and 2021, we explore whether QFII-invested FMCs (Q-FMCs) learn more from their domestic counterparts (D-FMCs) than other foreign-invested FMCs (NQ-FMCs). Our findings show that Q-FMC-managed mutual funds exhibit portfolio allocations more closely aligned with local DFMCs than those managed by NQ-FMCs. This imitation is particularly pronounced when selecting new stocks, enhancing portfolio performance, but not when rebalancing existing positions. Additionally, Q-FMCs trade more actively than NQ-FMCs. Robustness checks confirm these results across various ownership structures, fund characteristics, market conditions, and regulatory changes. These findings highlight the dual role of QFIIs as both investors and learners in China’s evolving financial landscape, offering insights into how foreign capital integrates into emerging mutual fund markets, informing regulatory policy aimed at fostering cross-border financial development.

We introduce a real-time dispersion measure based on cross-sectional stock returns explicitly designed to capture retail-driven speculative episodes. Elevated return dispersion effectively identifies periods characterized by intensified retail investor trading behaviors, driven by salience, diagnostic expectations, and extrapolative beliefs. During these high-dispersion states, momentum strategies collapse, and short-term reversals become dominant. Conditioning momentum strategies on our dispersion measure resolves the longstanding puzzle of missing momentum in retail-intensive markets such as China, substantially enhancing profitability. A dynamic rotation strategy between momentum and short-term reversal portfolios guided by dispersion states achieves annualized Sharpe ratios nearly double those of static approaches. Extending our analysis internationally, we employ Google search trends as proxies for retail investor attention, confirming that dispersion robustly predicts momentum and reversal returns globally. Our findings underscore the behavioral channel through which retail-driven speculation conditions momentum dynamics, providing clear implications for dynamic portfolio management strategies.

Against the backdrop of potential irrational trading behaviours in financial markets, this study investigates the V-shaped disposition effect in the selling activities of portfolios managed by securities investment funds in China. Utilising quarterly holdings data (2018–2024) of Chinese securities investment funds, alongside daily turnover rates and closing prices of their fund-heavy stocks listed in China's A-share market, a Fama-MacBeth regression analysis is conducted. The empirical results provide robust evidence of a significant V-shaped disposition effect in these fund investments, primarily driven by speculative trading. Moreover, this effect significantly and positively predicts future stock returns of Chinese A-shares. This study enhances understanding of institutional investors' trading behaviours—particularly mutual funds in China—and their decision-making processes in financial markets.

Using a novel and dynamic measure of fund-level herding that captures the tendency of a fund manager to imitate the trading decisions of the institutional crowd based on a sample of 3490 mutual funds in China for 21 years between 2003 and 2023, we find that funds with higher herding tendencies face significantly elevated delisting risks. Additionally, herding behavior is associated with shorter fund lifespans, smaller asset bases, and higher portfolio manager turnover rates. These results remain robust after employing a battery of methods to address endogeneity concerns. Collectively, our study demonstrates that herding substantially amplifies funds’ running risks.

This paper examines why tax incentives fail to stimulate participation in China's third-pillar commercial pension insurance, emphasizing the role of rational inattention. Using household survey data from China Family Panel Studies (CFPS) spanning 2014-2022 and a difference-in-differences-in-differences (DDD) design, we find that pilot policy generated a statistically insignificant average effect on participation, with rational inattention - proxied by financial literacy - explaining much of its ineffectiveness. We develop a dynamic consumption-portfolio model featuring costly information acquisition, and then resolve limitations of standard models through a dynamic framework with distinct savings channels and policy-focused rational inattention. The models show that rational inattention distorts perceptions of tax benefits and wage growth, raising participation costs, while multiple savings channels dilute incentives. Only households with higher financial literacy substantially respond to the policy. Our results reveal how cognitive frictions undermine pension reform and offer implications for designing behaviorally-informed retirement schemes.

We revisit cross-asset time-series momentum (XTSM) and single-asset time-series momentum (TSM) in China's stock and bond markets. With a fixed-effects model, we find a positive momentum from bonds to stocks and a negative momentum from stocks to bonds, with both momentum persisting for no more than six months. By employing a cross-grouping method, we find that the choice of lookback periods and asset signals impacts the performance of XTSM and TSM. A comparison between XTSM, TSM, and time-series historical (TSH) portfolios reveals that XTSM outperforms in small/midcap stocks and government bonds, while its performance is weak in large-cap stocks and corporate bonds. A spanning test confirms that XTSM generates excess returns that other pricing factors can not explain. XTSM is more prone to momentum crashes. Increased market stress has similarly adverse effects on XTSM and TSM. Furthermore, Market illiquidity, IPO counts, new investor accounts, and consumer confidence index positively correlate with the returns of XTSM and TSM portfolios, while IPO first-day return and turnover rate correlate negatively. The effects of these sentiment indicators exhibit heterogeneity.