The global AI inference market—reaching approximately $90–100 billion annually and growing at 18% CAGR—operates without organized exchange infrastructure. We document three market failures: resource misallocation (80% of China’s newly built compute capacity sits idle), price opacity (100-fold price dispersion across providers of equivalent quality), and unhedged risk exposure (85% of enterprises miss AI cost forecasts by more than 10%). Following the market design tradition of Roth [2002] and Budish et al. [2015], we propose the AI Service Right (ASR) as a transferable property right on AI compute and the AI Service Unit (ASU) as a quality-adjusted, cross-platform unit of account grounded in hedonic price theory [Rosen, 1974]. The ASU is modality-neutral: billing prices across text, image, video, and speech modalities are unified via eq-token conversion factors (κimg ≈ 2,667 eq-tokens per image; κvid ≈ 2,667 per second of video; κspc ≈ 7 per second of audio), and modality-appropriate benchmark sets (MMLU/HumanEval for language; FID/CLIP Score for image; FVD/CLIPSIM for video; MMBench for multimodal) supply the quality in dex via PCA. We design a hybrid secondary market architecture synthesizing mechanisms from four orthogonal market traditions: foreign exchange markets (cross-platform exchange rates and PPP-analog arbitrage via the ASU); equity markets (Central Limit Order Book, market making, clearing); electricity markets (Compute Locational Marginal Pricing for spatial scarcity signals); and decentralized finance (Automated Market Maker for long-tail liquidity). We establish nine formal propositions: bilateral trading is generically inefficient; Compute Locational Marginal Pricing decomposes nodal prices into system marginal cost, capacity congestion, and bandwidth premia; no-arbitrage equi librium holds with capital constraints (extending Shleifer and Vishny 1997); the ASR market Pareto-improves over bilateral trading; market prices are more in formative under ASR; the hybrid CLOB-AMM architecture weakly dominates either mechanism alone; platform adoption admits multiple equilibria with a coordination trap; financialization may improve or reduce price informativeness depending on speculator-hedger composition; and a hedonic micro-foundation justifies the ASU definition. Calibrated agent-based simulation (500 steps, 30 Monte Carlo runs) provides computational validation: the hybrid architecture reduces price dispersion by 90% relative to bilateral trading, and order-of-magnitude welfare estimates suggest enterprise procurement cost savings of 0.2–20% (net of ASR transaction costs; see Table 7) and potential TFP gains from compute reallocation of up to $29.9 billion annually. We propose a phased implementation roadmap from shadow ledger to full financialization, and we engage critically with the concern that financialization may not reduce intermediation costs [Philippon, 2015].

We investigate the capabilities of Large Language Models (LLMs) in generating novel alpha factors for option returns. Utilizing a structured prompt-engineering approach, LLMs like GPT-5 can directly create factors for two distinct options markets: the mature U.S. market and the emerging Chinese market. Empirical analysis further reveals that the LLM-generated factors exhibit remarkable and robust performance, delivering statistically signifcant returns in both all-sample and extensive out-of-sample tests. Beyond their statistical signifcance, such factors are economically meaningful. They display low self-correlation, indicating genuine innovation, and are grounded in sound economic rationale derived from market microstructure and behavioral fnance principles, showcasing a key advantage over traditional machine learning models.

How do market-design reforms affect secondary-market liquidity in an emerging equity market? We examine China’s ChiNext reform, which combined registration-based IPO issuance with changes in secondary-market trading rules. Using firm-month observations and high-frequency order-book data, we compare incumbent ChiNext firms with Main Board firms in a difference-in-differences framework. The results show that the reform significantly improves secondary-market liquidity, with consistent evidence across alternative DID estimators and liquidity measures. Dynamic estimates show comparable pre-reform movements and increasingly favorable liquidity outcomes after implementation. Cross-sectional analysis indicates stronger liquidity improvements among larger and more profitable firms. Microstructure results show narrower quoted spreads, broadly stable displayed depth, and a shift in order flow toward buyer-initiated trading. Overall, the findings demonstrate that the integrated ChiNext reform package improved secondary-market liquidity and reshaped the cross-sectional distribution of liquidity gains, highlighting the market-quality consequences of institutional reform in emerging equity markets.

This study integrates microstructure theory with asset pricing to investigates how the idiosyncratic volatility (IVOL) puzzle operates through specialized liquidity risk channels in China’s A-shares market. We employ intraday transactions data to perform a novel decomposition of liquidity into its variable (informational) and fixed (transitory) components. We show that the anomalous negative relationship between IVOL and future returns emerges from the intricate interaction of liquidity risk exposure, information and arbitrage constraints, and measurement biases. Specifically, the variable component tied to informed trading and adverse selection exposes high-IVOL stocks to greater arbitrage risk during liquidity shocks, while the fixed component exacerbates their vulnerability to short-term market-making cost fluctuations. Our results reveal that the IVOL puzzle is not a statistical artifact but a rational pricing phenomenon driven by omitted liquidity risk, mediated by the country’s unique institutional environment and monetary conditions.

This study examines the mean-reversion behavior of trading volume in China’s A-share market, with a focus on the speed at which abnormal surges dissipate. We compare two competing hypotheses: the stealth-trading hypothesis, where persistent volume reflects order-splitting by informed traders, and the informational-efficiency hypothesis, which interprets faster reversion as a sign of efficient information absorption. Using the Ornstein–Uhlenbeck (OU) model, we estimate the reversion speed for over 3,000 stocks and link it to firm- and industry-level characteristics. We find that trading volume is strongly mean-reverting, with over 98% of stocks classified as stationary. The OU model forecasts reversion speed with less than 7% error. Faster reversion is associated with larger size, higher analyst coverage, lower volatility, and greater liquidity. Notably, reversion speed increased after the 2006 IFRS reform but declined following Stock Connect, suggesting that stock market policies can influence informational efficiency. Our OU-based methodology offers a simple, observable proxy for monitoring how quickly markets process information. These results position trading volume as a core variable in market microstructure research and policy evaluation.

This study investigates the mean-reversion behavior of trading volume, using China’s A-share market as a representative setting characterized by dispersed retail investors, frequent public disclosures, and active policy interventions. We compare two competing interpretations:the stealth-trading hypothesis, in which persistent volume reflects order-splitting by informed investors, and the informational efficiency hypothesis, which links faster volume reversion to more effective information processing. Using the Ornstein–Uhlenbeck (OU) model, we estimate reversion speeds for over 3,000 stocks and relate these to firm- and industry-level characteristics. We find that trading volume is broadly mean-reverting, with over 98% of stocks exhibiting stationarity. The OU model forecasts reversion speed with less than 7% error. Faster reversion is associated with larger firm size, greater analyst coverage, lower volatility, and higher liquidity. Notably, reversion speed increased after accounting reforms but declined following capital access liberalization, suggesting that regulatory policy can both enhance and impair informational efficiency. These findings position reversion speed as an observable proxy for market responsiveness and highlight trading volume as a central variable in empirical market microstructure research.

The Hurst exponent is a key parameter for characterizing the long memory of high-frequency time series. However, traditional estimators often exhibit systematic biases due to the influence of high-frequency endogenous noise and low-frequency trend shifts. Theoretical derivations show that endogenous noise contemporaneously correlated with the latent signal possesses a spectral density in the first-differenced series that is asymptotically equivalent to a squared sine functional form. Accordingly, the proposed estimator incorporates a corresponding spectral density component to fit the high-frequency error. Simultaneously, the model introduces a SCAD penalty term to control the low-frequency spectral divergence caused by structural breaks, thereby mitigating spurious long memory in parameter estimation. Monte Carlo simulations demonstrate that the Penalized Mixture Whittle estimator yields smaller finite-sample biases and root mean square errors in scenarios involving both trend disturbances and endogenous noise. Empirical analysis shows that the estimates obtained using this method are robust to changes in sampling frequency. In further volatility forecasting experiments on commodity futures, the linear forecasting model constructed based on the parameter set achieves higher prediction accuracy than benchmark models such as HAR, as confirmed by the Diebold-Mariano test. This paper provides an effective econometric tool for high-frequency data inference in the presence of composite statistical disturbances.

This study identifies a significant settlement effect in China’s equity options market, where price decline and pre-settlement return momentum exists on the settlement Friday (T+2) due to a temporal misalignment between option expiration (T) and the T+1 trading rule for the underlying asset. We attribute this phenomenon to three distinct behavioral channels: closing pressure from put option unwinding, momentum-generating predatory trading by futures-spot arbitrageurs exploiting liquidity fragility, and an announcement effect that attenuates the anomaly by adjusting spot speculators' expectations. Robust empirical analysis identifies predatory trading as the primary driver of the settlement effect.These findings offer critical insights for market microstructure theory and the design of physically-delivered derivatives.

This paper examines bond trading activity and transaction cost differences between the bilateral Over-the-Counter (OTC) and the centralized Central Limit Order Book (CLOB) venues in the China interbank government bonds market, structured as all-to-all. Using a novel trade-level dataset, we estimate that CLOB reduces transaction costs by 0.66 basis points compared to OTC, highlighting the efficiency of its centralized trading mechanism. Furthermore, our analysis of cross-venue selection patterns reveals that the CLOB venue disproportionately facilitates core traders, orders with standardized sizes and settlement speeds, and newly issued bond trades. Despite CLOB’s cost advantages, the continued use of OTC is justified by its unique benefits, including mitigating information leakage, enabling designated counterparties, and facilitating position rebalancing. These findings offer insights into how market microstructure and trading mechanism affect asset liquidity.

We introduce a novel perspective on private information, specifically microstructure-based private information, to unravel how institutional investors predict stock returns. Using tick-by-tick transaction data from the Chinese stock market, we find that in retail-dominated markets, institutional investors positively predict stock returns, consistent with findings from institution-dominated markets. However, in contrast to the traditional view that institutional investors primarily rely on value-based private information, our results indicate that microstructure-based private information contributes almost as much to their predictive power as value-based private information does, with both components jointly accounting for approximately two-thirds of the total predictive power of institutional order flow. This finding reveals that retail investors’ trading activities significantly impact institutional investors, naturally forcing them to balance firm value information with microstructure information, thus profoundly influencing the price discovery process in the stock market.