Historical Momentum-Index Constituents by Rebalance

Turn recorded ten-stock momentum baskets into a rebalance panel with complete ranks, additions, removals, and turnover, without rebuilding history from revised inputs.

Historical momentum-index constituents should be studied as a sequence of recorded baskets. Alphanume's Quant Galore Momentum Index dataset provides the maintained ten-stock monthly basket with the observation date, ticker, and rank within the basket. Rank one is the strongest included momentum name, while rank ten is the lowest-ranked constituent still selected.

The Momentum Index guide explains the basket workflow. The historical panel supports turnover, rank persistence, concentration, and post-rebalance analysis. It does not make historical compounding an expected return, and it does not remove the costs or risks of holding only ten stocks.

Build the panel from recorded basket dates

Request a date range, group rows by the served date, and treat those groups as recorded baskets. Do not generate assumed month-end dates or force a fixed calendar interval. Each valid rebalance group should have ten unique symbols and unique ranks from one through ten.

The small source schema is sufficient because turnover is derived by comparing adjacent baskets. Preserve source values and place derived fields in separate columns. That distinction makes it easy to recheck ranks against the API while changing the turnover definition independently.

Retrieve all baskets and validate them

Use logically valid date bounds for the intended research window. Save the raw response alongside the derived panel. For each date, run invariant checks before calculating returns. A missing rank or duplicate symbol should stop the pipeline rather than being filled from a neighboring month.

GET /v1/quant-galore-momentum-index
date_gte=2023-01-01
date_lte=2025-12-31

for each served date:
  assert count(rows) == 10
  assert count(unique(ticker)) == 10
  assert sort(rank) == [1,2,3,4,5,6,7,8,9,10]

for each adjacent basket pair:
  additions = current tickers minus prior tickers
  removals = prior tickers minus current tickers
  retained = intersection of both sets

The dashboard documentation states that the monthly basket updates at 4:05 PM America/New_York on rebalance. A return study must choose a tradable time after that availability point. Same-close execution is not implied by the constituent record and requires a separately supported price and implementation assumption.

Measure turnover and rank movement

Report both name turnover and estimated trade turnover. Name turnover counts additions and removals. Trade turnover also includes weight resets among retained names if the portfolio returns to equal weights. For retained tickers, calculate rank changes, but remember that rank is ordinal and relative within the ten selected names.

• Show additions, removals, and retained names for every transition.
• Keep delisted or renamed symbols in the outcome ledger.
• Apply transaction costs to entries, exits, and weight resets.
• Report issuer and sector contributions to reveal concentration.

A ten-stock basket can be dominated by one constituent or one industry. Publish constituent-level attribution alongside aggregate returns. Rank one is not a calibrated return forecast, and the distance between ranks one and two does not represent a fixed expected-return gap.

Turnover can be expressed several ways, so define it before comparing periods. A simple name-turnover rate divides additions by ten. A portfolio-turnover estimate uses the absolute weight changes required to move from the pre-rebalance portfolio to the target basket. These values differ when retained names drift away from equal weights. Store both when the research addresses implementation rather than only membership stability.

The panel can also measure persistence without using returns. Calculate how often rank-one names remain in the next basket, how far retained names move in rank, and how long constituents remain selected. These descriptive statistics reveal basket stability, but they should not be mined for a holding rule and then presented as an independent test. Reserve later dates or use a declared validation design.

Avoid revised reconstruction and compounding claims

The main failure mode is reconstructing an old basket with today's universe, corrected prices, or a different momentum formula and calling it the maintained index. Such a reconstruction can be useful as an independent factor study, but it is a separate portfolio. Label the recorded and reconstructed series distinctly.

The second failure is showing gross compounding without turnover, slippage, concentration, or missing-security treatment. Predefine the benchmark, availability lag, weight convention, and corporate-action handling. Report gross and cost-adjusted performance, then show how much each rebalance contributes.

Audit one pair of rebalances

As a concrete next action, query two consecutive recorded dates, validate both ten-name rank sets, and export their additions, removals, retained names, and rank changes. Attach prices under a written post-publication timing rule and calculate equal-weight turnover plus a conservative cost estimate. That two-basket audit verifies the panel construction before a long historical compounding study.