China A-Share Factor Investing: A 2026 Data Guide
Why size, value and short-term reversal historically pay in China A-shares while price momentum is weak, grounded in academic consensus, plus live SurgeFlow CN factor data and a free API.
Category: Factor Research
What pays in China A-shares, and what does not
In China A-shares, academic work documents strong size and value premia and a notable short-term reversal effect, while conventional price momentum is weak or absent (Liu, Stambaugh and Yuan, 2019). SurgeFlow does not assert these as proprietary results; it publishes daily, point-in-time CN factor returns you can inspect on /factors and pull via the free API.
The A-share market differs structurally from developed markets. It is retail-dominated, with high turnover and frequent reversals, and the smallest listed firms have historically carried an embedded reverse-merger or shell-value premium that contaminates a naive size factor. Because of this, the established literature argues for screening out the smallest decile of stocks before constructing a China size factor, and for replacing the standard book-to-market value proxy with an earnings-to-price proxy, which the authors find better captures the value effect in this market (Liu, Stambaugh and Yuan, 2019, Journal of Financial Economics).
The practical implication is that factor results imported wholesale from US research should not be assumed to hold in China. SurgeFlow treats each market separately: the CN factor models are estimated on A-share data, benchmarked against the CSI 300, and gated by the same publication rules used across all four markets rather than by a global default.
How the factors are constructed and tested
SurgeFlow builds Fama-French style factors for China A-shares: SMB (size), HML (value), RMW (profitability), CMA (investment), WML (momentum) and a LIQ liquidity factor. Each factor is a daily return series formed from long and short legs of ranked portfolios, with point-in-time fundamentals where they exist. For China that is partial: book value and size are formed from filing-date data, while profitability uses a latest-state proxy — so a historical profitability signal can carry information that was not knowable on its formation date.
Factor relevance is assessed with standard cross-sectional methods. Fama-MacBeth regressions estimate the average premium attached to each characteristic by running a cross-sectional regression of returns on factor loadings each period and averaging the resulting slopes over time, with t-statistics computed on that time series. The GRS test (Gibbons, Ross and Shanken, 1989) checks whether a candidate factor model leaves no economically meaningful alpha across test portfolios, and VIF diagnostics flag multicollinearity so that overlapping factors are not double-counted. An effective-direction resolver records the sign in which a factor actually pays in the data, rather than assuming the textbook direction. - SMB size, HML value, RMW profitability, CMA investment, WML momentum, LIQ liquidity — daily return series - Point-in-time book and size inputs; profitability uses a latest-state proxy, and survivor bias is labelled rather than removed - Fama-MacBeth premia, GRS model test, VIF multicollinearity checks, minimum-observation gates - A publication gate that blocks factors which fail integrity or significance checks from being presented as usable
Sleeves, weighting and the composite score
Factor Backtest V2 turns single factors and combinations into long sleeves and reports comparable metrics per sleeve: Sharpe (excess), annualized Information Ratio, maximum drawdown, turnover and the number of observations. Sleeves are built under two weighting schemes. Equal weight (EW) gives each name the same allocation. Risk parity (RP) instead scales positions so each holding contributes a similar share of portfolio risk, which typically dampens the influence of the most volatile names.
To pick a representative winner without cherry-picking, SurgeFlow ranks sleeves on a z-score composite, argmax of 0.5 times z(Sharpe) plus 0.5 times z(Information Ratio), so a sleeve must score well on both risk-adjusted return and consistency of excess return. As a concrete cross-market reference point, the current US composite winner is the WML+RMW Top-30 RP sleeve over roughly ten years, ranked first by both Sharpe and Information Ratio (SurgeFlow's Phase 1 backtest is research_backtest_insample — in-sample factor-attribution, not live out-of-sample evidence). That is a US result; the CN sleeve rankings are computed on A-share data and may look different, given the documented weakness of price momentum in China. The honest move is to read the live CN values rather than assume the US pattern transfers.
What SurgeFlow data this uses
The China A-share factor models are estimated on Tushare-sourced A-share price and fundamental data, benchmarked against the CSI 300. Factor return series and per-sleeve backtest metrics are published on the live factor and backtesting pages, subject to the publication gate. Formation is month-end rather than daily — the CN factor-signal series steps 2026-05-29, 2026-06-30, 2026-07-31 — and the gate records a point-in-time state and a minimum-observation count on every row. It does not certify survivorship: the backtest engine labels its own runs in-sample and carries a survivor-bias label forward from the underlying factors rather than removing the bias. Read what is shown as a research record, not a track record.
We deliberately do not print a fixed China size, value or momentum premium, Sharpe or weight on this page. Those values move with the data and are best read live. Coverage detail for the CN market — universe, sources and known gaps — is documented on the CN equity market data coverage page, and the same factor methodology applied to other markets is described on the US and JP coverage pages for comparison. - Source: Tushare A-share prices and fundamentals; benchmark CSI 300 - Live factor returns on /factors; per-sleeve metrics on /backtesting - Coverage, universe and gaps detailed at /research/cn-equity-market-data-coverage
Caveats and cross-market framing
Two cautions matter. First, the China A-share market has changed over the sample: regulatory shifts, the gradual opening to foreign capital and changes in the small-cap shell premium mean factor behaviour is not stationary, so a premium documented in one period need not persist. Second, the academic findings cited here are public consensus from peer-reviewed work, not SurgeFlow proprietary discoveries; readers can and should verify them against the original papers.
Nothing here is investment advice, a recommendation, or a prediction. It is a description of research data and the methods used to produce it. The strongest claim we make is methodological: factors are estimated per market, tested with standard statistical procedures, and withheld when they fail. The numbers themselves live in the data.
Get the data: free API key
You can inspect the live China A-share factor returns on /factors and the per-sleeve backtest metrics on /backtesting, then pull the same data programmatically. The free public API covers all four markets at 2,000 requests per day, with realtime, screen and hotlist endpoints per market plus a summary endpoint; the schema is published at /api/public/openapi.json.
To start, create a free API key at /account and open the live console at /membership/api. From there you can query the CN factor and market endpoints directly, compare China against the US and JP factor pages, and check the documented coverage before relying on any series.
Sources and further reading
Primary, academic, and SurgeFlow data references for the claims above: - Liu, J., Stambaugh, R. F., and Yuan, Y. (2019). Size and value in China. Journal of Financial Economics, 134(1), 48-69. - Fama, E. F., and French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics, 33(1), 3-56. - Fama, E. F., and French, K. R. (2015). A five-factor asset pricing model. Journal of Financial Economics, 116(1), 1-22. - Fama, E. F., and MacBeth, J. D. (1973). Risk, return, and equilibrium: empirical tests. Journal of Political Economy, 81(3), 607-636. - Gibbons, M. R., Ross, S. A., and Shanken, J. (1989). A test of the efficiency of a given portfolio. Econometrica, 57(5), 1121-1152.