Factor surfaces are beta and return the SECAPI-Maturity: beta header. Validate them for your use before treating a result as production decision input.
2015-01-01 history floor. A target is not row-level proof; use the returned history and freshness metadata for the individual factor and period.
Construction of stock-basket factors
Universe
U.S. market, style, sector, and industry baskets use a monthly refreshed, point-in-time U.S. equity universe. The documented screen requires at least 100Mmarketcapitalization,5M average daily dollar volume, and a $1.00 share price. It excludes warrants, units, rights, ETFs, ETNs, and ADR certificates, and includes NYSE, NASDAQ, AMEX, ARCA, and BATS listings.
Historical reference snapshots determine membership at rebalance to avoid look-ahead bias. When those snapshots are absent, market capitalization and SIC classification can use the latest reference data. That fallback is approximate rather than point-in-time.
Characteristics and returns
Style factors are cap-weighted long-short decile spreads: securities are ranked on a characteristic, assigned to deciles, and the top-minus-bottom return is rebalanced monthly. Definitions can use point-in-time SEC XBRL fundamentals, SIC classification, or documented basket and proxy inputs.
ETF-backed thematic factors use the ETF return less a relevant benchmark. Curated thematic baskets use equal-weight constituent returns less a relevant benchmark. Macro and country factors use a documented proxy spread. The catalog and methodology metadata identify what applies to a specific factor; no single construction applies to every category.
Purification and scaling
Factors have one to three declared parents. SEC API uses a rolling 156-trading-day OLS regression to estimate residual return not explained by those parents. The deliberately limited hierarchy keeps the relationship inspectable and preserves the defining basket signal.
For residual scaling, the model uses the preceding 60 residual returns plus the current residual to estimate sample volatility, annualizes with 252 trading days, targets 10% annualized volatility, and caps leverage at 3.0. Returns are then converted to rolling z-scores for comparison. These are model outputs, not forecasts, performance claims, or investment recommendations.
Interpretability requirements
During U.S. market hours, intraday snapshots can include raw, purified, scaled, and z-score returns only when the required inputs are current. For every result that informs a model, report, or decision, retain request identifiers, methodology, source-rights, freshness, materialization, and degradation metadata when supplied.
Read Factor provenance and Factor freshness alongside any factor output.
