Historical Market Cap: A Practical Guide

The Number That Changes Everything Else

Market capitalization is the single most referenced fundamental metric in equity research. It determines index membership, factor classification, liquidity expectations, institutional eligibility, and position sizing. Nearly every systematic strategy uses it somewhere in the pipeline.

But here's the problem: most data sources give you today's market cap. If you're building a backtest and you need to know what Apple's market cap was on March 15, 2019, you need the value as it was known on that date — not a number recalculated with today's share count or adjusted for subsequent splits and issuances.

The Historical Market Cap dataset provides exactly this: point-in-time market capitalization and shares outstanding for U.S. equities, as they were known on each historical date. No retroactive adjustments, no forward leakage.

Why Point-in-Time Matters

Imagine you're backtesting a strategy that only trades stocks above $1 billion market cap. You're screening the universe on January 1, 2020, and a stock shows $1.2 billion. Looks like it qualifies.

But that $1.2 billion is today's value, back-filled into the historical record. On January 1, 2020, the actual market cap was $600 million. Your strategy should never have traded it. You've introduced lookahead bias — using information from the future to make decisions in the past.

This is not a theoretical concern. Shares outstanding change frequently due to buybacks, issuances, stock-based compensation, and dilution. A company that has 100 million shares today might have had 80 million shares three years ago. Multiply by the historical price and you get two very different market cap numbers.

Point-in-time data eliminates this problem. Each record reflects the shares outstanding and market cap as they were known on that specific date. Your backtest sees what you would have seen if you were actually trading on that day.

What the Dataset Provides

For each equity on each date, you get two fields:

market_cap: The market capitalization as it was known on that date. This is computed from the shares outstanding and closing price at the time.

shares_outstanding: The number of shares outstanding as reported at that point in time. This is the raw share count before any subsequent changes from buybacks, issuances, splits, or other corporate actions.

These two fields together give you a complete picture of a company's equity size at any historical moment.

How Traders and Researchers Use This

Size factor construction. Market cap is the standard proxy for the size factor in multi-factor models. If you're building SMB (small minus big) or any size-based portfolio, you need historical market caps that reflect what was known at the time. Using current values contaminates the factor with lookahead bias.

Universe eligibility screening. Most systematic strategies define a tradable universe using market cap thresholds: "trade only stocks above $500M," or "focus on micro-caps below $300M." Point-in-time market cap lets you apply these filters correctly at each historical date.

Signal normalization. Many trading signals are more meaningful when normalized by size. Dollar volume, notional positioning, options flow — these all look different for a $2 billion company vs. a $200 billion company. Historical market cap gives you the denominator you need.

Dilution and buyback monitoring. Changes in shares outstanding over time tell you about structural capital flows. A steadily declining share count suggests aggressive buybacks. A rising share count signals issuance or dilution. Tracking this historically can be a signal in its own right.

Position sizing. Some position sizing frameworks scale inversely with market cap (more conviction in larger names) or directly (larger positions in smaller names for more impact). Historical market cap ensures your sizing logic was consistent with the data available at the time.

What the Data Looks Like

{
"date": "2026-02-06",
"ticker": "AAPL",
"market_cap": 4109599296360,
"shares_outstanding": 14776353000
}

Clean, precise, and ready to plug into any pipeline. The dataset supports pagination for efficient retrieval of large date ranges or full-universe pulls.

Key Details

The Bottom Line

Market cap is the most basic building block in equity research, and getting it wrong poisons everything downstream. If your backtest uses current market cap values for historical dates, your universe construction is wrong, your factor exposures are wrong, and your performance numbers are unreliable.

The Historical Market Cap dataset gives you the correct number for every stock on every date — as it was actually known at the time. It's not the most exciting dataset in the catalog, but it might be the most important. Everything else you build sits on top of it.