US vs Japan Market Anomalies: 2026 Data Study

A primary-source comparison of how documented equity anomalies differ between the US and Japan, with academic attribution and links to live cross-market factor data.

Category: Factor Research

How do US and Japan market anomalies differ? US and Japan share the size, value, and profitability anomalies documented by Fama and French, but the momentum anomaly differs sharply: Asness (2011) and Fama-French (2012) show momentum is robust in the US yet historically weak or absent in Japan, while value has been comparatively strong in Japan. SurgeFlow publishes live factor returns for both markets.

An equity anomaly is a documented pattern of average returns that a single-factor market model fails to explain. The standard academic toolkit for measuring these is the Fama-French factor framework: SMB (size), HML (value), RMW (profitability), CMA (investment), plus a price momentum factor (WML / UMD) and a liquidity factor (LIQ). The interesting research question is not whether these anomalies exist somewhere, but whether the same factor earns a premium of the same sign and significance in two structurally different markets.

The US is the most heavily studied equity market and the original setting for the three- and five-factor models (Fama and French, 1993; 2015). Japan (TOPIX) is the most important out-of-sample test because it is large, liquid, and developed, yet historically behaves differently on at least one core factor. - Value (HML): documented in both markets; Fama and French (2012) report value premia across developed regions including Japan, where it has been comparatively pronounced. - Momentum (WML/UMD): strong and persistent in the US; historically weak or statistically insignificant in Japan (Fama and French, 2012; Asness, 2011 'Momentum in Japan: The Exception That Proves the Rule', JPM). - Size (SMB): present in both, though the size premium is widely noted as time-varying and weaker post-publication in the US. - Profitability and investment (RMW, CMA): added in the five-factor model (Fama and French, 2015); evidence is broadly supportive internationally but more recent and less settled than value.

Why Japan is the canonical out-of-sample test for momentum

Momentum is one of the most robust anomalies in US data, but Asness (2011) framed Japan as 'the exception that proves the rule.' Stand-alone momentum has been historically weak in Japan, yet Asness argued that a combined value-and-momentum strategy still works there because value and momentum are negatively correlated, so a portfolio that holds both is more diversified than either alone. The Japanese case is therefore not evidence that momentum is spurious, but a reminder that single-factor premia are regime- and region-dependent.

This matters for anyone building cross-market factor models. A US-tuned momentum sleeve cannot be assumed to transfer to Japan; the honest approach is to estimate each factor separately per market and report the realised premium rather than borrow a US result. SurgeFlow's design reflects this: factors are constructed and published per market, and a publication gate blocks factors that cannot be supported by the available data for a given market.

Methodology: how anomalies are measured

Anomalies are tested with a small set of well-established methods, all of which SurgeFlow's factor system follows.

Factor construction sorts the cross-section into portfolios (for example, large-minus-small for SMB, high-minus-low book-to-market for HML) and computes the daily long-short return spread. To judge whether a set of factor portfolios is jointly priced, the GRS test (Gibbons, Ross and Shanken, 1989) checks whether the intercepts (alphas) of test-asset regressions are jointly zero; rejecting that null means the model leaves an unexplained anomaly. Cross-sectional pricing of a characteristic is tested with the Fama-MacBeth (1973) two-pass regression, which estimates a time series of cross-sectional slopes and tests their average against zero.

For portfolio backtests, SurgeFlow's Factor Backtest V2 builds long sleeves of single and combined factors under two weighting schemes: equal weight (EW) and risk parity (RP), where RP scales each position inversely to its volatility so that risk contributions are balanced rather than dollar exposures. Each sleeve reports excess Sharpe, annualized information ratio, max drawdown, turnover, and number of observations, behind data-quality gates for point-in-time data, survivorship handling, and a minimum-observations floor; the GRS test and other diagnostics are reported alongside, not used to gate or exclude sleeves (the current Phase 1 run is in-sample, research_backtest_insample). Sleeves are compared using a z-score composite of Sharpe and Information Ratio, with current figures available on /factors and /backtesting. This page does not declare a winning sleeve.

What SurgeFlow data this uses

This comparison draws on SurgeFlow's live, multi-market equity factor system, which covers all four markets including the US and Japan. US fundamentals and prices come from FMP, with SEC EDGAR 13F for institutional ownership. Japan (TOPIX) is built from J-Quants prices and fundamentals plus EDINET large-holding reports for major-shareholder ownership.

For both markets SurgeFlow computes Fama-French style factors (SMB, HML, RMW, CMA, WML, LIQ) with daily factor returns, an effective-direction resolver, VIF multicollinearity diagnostics, and a publication gate that blocks factors which cannot be honestly supported. The realised premia, signs, and significance for each factor are not asserted here as fixed numbers; they are published live and change over time. To read the current US and Japan values rather than a snapshot, see the live models on /factors and the per-sleeve backtest metrics on /backtesting. This page is research data, not investment advice.

Honest cross-market caveats

Anomaly research carries well-known hazards, and they apply equally to the US and Japan. Publication and data-mining bias mean some premia shrink or vanish after they are documented; the size effect is a frequently cited example. Survivorship and look-ahead bias can manufacture premia that did not exist in real time, which is why point-in-time data and survivorship handling are integrity gates rather than optional steps.

Cross-market comparison adds further caveats. Japan's coverage history, accounting conventions, and ownership-disclosure regime differ from the US, so a difference in a measured premium can reflect data structure as much as market behaviour. Currency, market microstructure, and differing sample lengths all affect comparability. The defensible position is to report what each market's own data shows, attribute every academic finding to its source, and let readers inspect the live figures rather than trust a single summary statistic.

See the live US and Japan factor data

The anomalies above are documented findings from the academic literature, attributed to their authors. The point of this page is to send you to the live, current evidence rather than a frozen number. Create a free API key at /account to query SurgeFlow's open multi-market data console: the free tier allows 2,000 requests per day across all four markets, including the US and Japan factor returns. From there you can pull current factor values, compare sleeves on /backtesting, and read methodology detail on /factors. All outputs are research data, not investment advice.

Sources and further reading

Primary, academic, and SurgeFlow data references for the claims above: - Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics. - Fama, E. F., & French, K. R. (2012). Size, value, and momentum in international stock returns. Journal of Financial Economics. - Fama, E. F., & French, K. R. (2015). A five-factor asset pricing model. Journal of Financial Economics. - Asness, C. (2011). Momentum in Japan: The Exception That Proves the Rule. Journal of Portfolio Management. - Fama, E. F., & MacBeth, J. D. (1973). Risk, return, and equilibrium: Empirical tests. Journal of Political Economy. - Gibbons, M. R., Ross, S. A., & Shanken, J. (1989). A test of the efficiency of a given portfolio. Econometrica. - Liu, J., Stambaugh, R. F., & Yuan, Y. (2019). Size and value in China. Journal of Financial Economics.