Factor Momentum in China's A-share Market: Evidence from 2000–2023

Factor Momentum in China's A-Share Market: Empirical Evidence (2000–2023)

Factor momentum strategies in China's A-share market have delivered annualized returns of approximately 9.91% with a Sharpe ratio of around 1.15; nearly double the performance of comparable strategies in the US market. That gap is not a coincidence, and it is not simply explained by an emerging market risk premium. It is rooted in the structural features that define Chinese equity markets: retail investor dominance, constrained short selling, daily price limits, and the sentiment-driven factor persistence those conditions produce.

This article draws on research by Gu, Xiong, and Chen (2024, China Journal of Econometrics) and the quantitative benchmarks established by Ma, Liao, and Jiang (SSRN 4148445) to systematically analyze factor momentum in China's A-share market from 2000 through 2023. The central finding: A-share factor momentum is a behavioral anomaly driven by sentiment, not a conventional risk premium; and because of short-sale constraints, the long leg has historically contributed more than 80% of total excess returns.

Factor Momentum: A Brief Global Context

Factor momentum refers to the tendency of recently outperforming investment factors (value, momentum, quality, and others) to continue outperforming. Unlike stock-level momentum, factor momentum operates across entire factor strategies, going long recently winning factors and short recently losing ones.

Quant Decoded's original backtest (see Factor Momentum Across Asset Classes: An Original Backtest) found that in US equity markets, a 1-month lookback factor momentum strategy produced annualized returns of approximately 7.2% with a Sharpe ratio of around 0.61; cross-asset diversification pushed the Sharpe ratio further to approximately 0.83. That evidence is compelling, but it is entirely drawn from developed markets dominated by institutional investors with functional short-selling infrastructure.

Applying the same framework to China's A-shares produces a systematic deviation, and the deviation runs in the stronger direction, not the weaker. Understanding why requires a close look at A-share market structure.

The Structural Distinctiveness of China's A-Share Market

Chinese equity markets differ from developed markets along several dimensions that together create the institutional conditions for unusually strong factor momentum.

Retail dominance has two reinforcing effects. First, retail investors update their beliefs about macroeconomic and factor-level information more slowly than institutional investors, creating meaningful lags in the transmission of factor signals. Second, retail herding pushes established factor trends beyond fundamental justification, extending factor persistence through behavioral amplification rather than rational repricing.

Short-sale constraints compound this dynamic. In developed markets, institutional arbitrageurs can short underperforming factor portfolios and accelerate the correction of mispricings. In A-shares, that correction mechanism was essentially absent before 2010 and remains constrained today. Factor-level mispricings therefore persist longer, and when they do correct, the adjustment occurs almost entirely through the long leg.

Price limits add a further dimension. The ±10% daily cap reduces extreme volatility, but it also slows price discovery, stretching factor return persistence across time in a way that mechanically supports momentum strategies.

Data and Methodology

The analysis covers the full A-share universe (all-A market), excluding ST-designated stocks and shares listed for fewer than 12 months. The sample runs from January 2000 through December 2023, providing approximately 288 monthly observations.

Five canonical factors are tested:

The factor momentum strategy ranks all five factors by their cumulative total return over the prior L months, going long the top-ranked factor and short the bottom-ranked factor each month. The primary lookback periods tested are 1 month and 12 months. Both time-series factor momentum (based on each factor's own absolute past return direction) and cross-sectional factor momentum (based on relative rankings across factors) are evaluated. The benchmark is the CSI 300 Total Return Index.

Empirical Results

Performance by Lookback Period

All four specifications are statistically significant (t-statistics above 2.0) and economically material. Cross-sectional factor momentum at the 1-month lookback is the strongest specification, with a Sharpe ratio of approximately 1.15, well above the US benchmark of approximately 0.61.

A-Share vs. US Factor Momentum: A Direct Comparison

The A-share advantage is not just about higher absolute returns; the gap in risk-adjusted performance is larger still. Annualized volatility is lower in China (~8.60% vs. ~11.80%), and the maximum drawdown is shallower (~-12.4% vs. ~-18.3%). The same factor rotation framework produces a materially superior risk-return profile in A-shares than in US equities.

Sentiment as the Driving Mechanism

Gu et al. (2024) identify the most important mechanistic feature of A-share factor momentum: strategy returns are strongly negatively correlated with investor sentiment. Factor momentum is most profitable precisely when market sentiment is depressed.

This pattern directly contradicts a risk-compensation explanation. Under a risk premium framework, returns should be higher in adverse conditions because risk is elevated, not because behavioral mispricings are larger. In A-shares, the mechanism is different: during low-sentiment periods, retail panic and belief rigidity create the most severe factor-level mispricings, while short-sale constraints prevent institutional arbitrageurs from correcting them quickly. Factor persistence therefore peaks during market stress.

The long-leg concentration reinforces this interpretation. In the full long-short strategy, the long leg has historically contributed more than 80% of total excess returns. Going long recently winning factors is effective; shorting recently losing factors is constrained by trading rules and delivers only marginal contribution.

Practical Considerations

A-share factor momentum has demonstrated strong persistence in historical data, but several dimensions warrant attention in practice.

On lookback period selection: the 12-month lookback generates lower turnover than the 1-month specification, and historically retains a higher proportion of gross returns net of costs; it may suit cost-sensitive investors better. The 1-month lookback produces a higher Sharpe ratio in gross terms, but the friction costs of frequent rebalancing erode a meaningful share of that advantage.

On market segmentation: CSI 300 large-cap constituents carry lower transaction costs than small and mid-cap stocks, and factor momentum has historically retained a higher net-of-cost return in the large-cap segment. Smaller stocks display stronger factor persistence, but liquidity constraints and market impact costs compress the alpha that is practically extractable.

On monitoring signals: sentiment indices (such as composite investor sentiment measures), margin loan balances, and turnover rates have historically shown some correlation with strategy returns; low-sentiment periods have tended to be relatively favorable for factor momentum.

On factor crowding: the rapid growth of Chinese quantitative private funds since 2019 has made factor crowding a structural risk that cannot be ignored. As more capital chases the same factor signals, some portion of the historical excess return has likely compressed.

Limitations

The main limitations of this analysis are as follows. First, the factor return series referenced here are constructed from academic data and differ from the actual mechanics of investable vehicles such as quantitative private funds and ETFs. Second, A-share data access for offshore investors is restricted, and the true cost of replication is likely higher than academic estimates suggest. Third, quantitative crowding may cause historical excess returns to decay going forward; this risk is particularly pronounced in China given the pace of institutional capital growth. Fourth, the Gu et al. (2024) analysis ends its sample in 2023, and post-2023 market dynamics have yet to be validated by subsequent research.