Center for Research in Security Prices (CRSP)

The Center for Research in Security Prices (CRSP) is a financial-data business known for research-quality histories of U.S. securities, returns, corporate events, and market indexes. CRSP began at the University of Chicago in 1960, and its main U.S. stock database now includes daily and monthly history extending back to December 1925. Morningstar acquired CRSP in February 2026.

CRSP research databases and CRSP market indexes are different products. The research databases retain the CRSP identity, while the former CRSP Market Indexes have been integrated and rebranded within Morningstar Indexes.

Key Takeaways

• CRSP provides licensed data, not investment recommendations or guaranteed error-free records.
• Its permanent identifiers help researchers follow securities and companies through ticker and name changes.
• Proper delisting returns, distributions, and corporate-action adjustments can materially change a backtest.
• Researchers must specify the database vintage, file format, filters, return variable, and sample dates.
• CRSP’s historical indexes should not be confused with the commercial index family that was rebranded after Morningstar’s acquisition.
• Long historical coverage reduces some data problems but does not remove survivorship, look-ahead, selection, or model risk.

CRSP History and Current Structure

CRSP was established to support empirical research into U.S. stock returns. The early project produced machine-readable security histories that allowed researchers to examine long periods systematically rather than assemble prices manually.

The organization expanded into multiple security, fund, Treasury, real-estate, and linked accounting databases. Morningstar completed its acquisition of CRSP from the University of Chicago on February 2, 2026. The current structure separates two related ideas:

A paper citing “CRSP” may therefore mean security-level research data, a historical research index, or a formerly CRSP-branded investable benchmark. The citation and methodology should identify which one.

Major Research Databases

Coverage, fields, start dates, and access rights differ by product. A statement that “CRSP starts in 1925” should not be applied to every database or every exchange.

Why Permanent Identifiers Matter

Ticker symbols are labels, not stable keys. A ticker can change, be reused by another company, or identify different share classes over time.

CRSP’s PERMNO is a permanent security identifier designed to follow one security through name and ticker changes. PERMCO identifies the company-level entity, which may have multiple securities.

Worked Example

Assume a security trades under ticker ABC for five years and changes its ticker to XYZ after a corporate name change. A ticker-based merge can mistakenly create two return histories or link the old ABC observations to a later unrelated company that reuses the symbol.

Using the same PERMNO before and after the name change preserves the security history:

The identifiers improve continuity, but the researcher still must handle mergers, multiple share classes, reorganizations, and company-level aggregation correctly.

Returns and Delistings

A security that disappears after a severe decline should not simply vanish from a sample. When both a regular-period return and a separate delisting return are available, the combined return is generally expressed as:

If a stock loses 20% during its final trading period and then has a -50% delisting return:

Dropping the delisted security or recording only the first -20% would materially overstate performance. Simply adding -20% and -50% would also be wrong because sequential returns compound.

The exact variable treatment, missing-value policy, and timing must follow the relevant CRSP guide and research design.

Legacy and Modern File Formats

CRSP released its newer CIZ flat-file format in 2022. It incorporates expanded underlying data and changes how monthly stock and index returns are derived from daily observations. Older research commonly uses the legacy FIZ format.

This creates an important reproducibility issue:

• the same broad sample can produce small differences across data vintages or formats;
• historical corrections can alter past observations;
• variable names and event treatment may differ; and
• code written for one format may not reproduce another study without mapping.

A credible analysis should record the product, data cut, format, extraction date, and program used.

Common Research Uses

Asset-pricing tests

Researchers use CRSP returns and market-cap data to form portfolios, estimate factor exposures, and test relationships between characteristics and returns.

Event studies

Security histories can be aligned with earnings announcements, mergers, or regulatory events to estimate abnormal returns. Event dates and benchmark-model choices remain separate sources of error.

Fund analysis

The survivor-bias-free mutual fund database helps include funds that merged or liquidated rather than studying only current survivors.

Backtesting

Investment practitioners can test historical rules, but CRSP does not make a backtest free of look-ahead bias. The code must use information available at each historical date and realistic rebalance, transaction-cost, and liquidity assumptions.

How to Design a Defensible CRSP Study

1. Define the eligible exchanges, share codes, security types, and date range.
2. Choose daily or monthly data and document the file format.
3. Use stable identifiers rather than ticker symbols as database keys.
4. Decide how to treat distributions, missing returns, suspensions, and delistings.
5. Use lagged market capitalization or accounting data where the design requires point-in-time information.
6. Include securities that later failed when they met the historical selection rule.
7. Document data revisions, merge logic, and the exact CRSP data cut.
8. Test whether conclusions survive alternative filters and reasonable transaction costs.

Risks and Limitations

• License restrictions: Redistribution and publication of licensed data are governed by the subscription terms.
• Vintage risk: Corrections and updated source data can change historical values.
• Coverage differences: Start dates and fields vary by exchange and product.
• Identifier misuse: PERMNO and PERMCO solve different security-level and company-level problems.
• Delisting bias: Missing or mishandled terminal returns can overstate results.
• Look-ahead bias: A current constituent list or later-known accounting value can contaminate a historical test.
• Data-mining risk: Testing many strategies increases the chance of finding accidental patterns.
• Implementation gap: A research return can exclude market impact, fees, taxes, borrowing constraints, and investability limits.

Sources

• Morningstar: CRSP Research Data Products
• CRSP U.S. Stock and Indexes Database Guide
• Morningstar acquisition of CRSP
• SEC filing confirming the 2026 CRSP acquisition

Related Terms

• Survivorship Bias: Distortion caused by retaining only securities or funds that survive.
• Delisting: Removal from an exchange, which can carry a separate terminal return.
• Market Capitalization: A common input to value-weighted indexes and portfolio sorts.
• Total Return: Price movement plus income under a defined calculation.
• Benchmark Index: A comparison standard, distinct from a security-level database.

FAQs

This page is educational and does not provide personalized investment, data-licensing, statistical, or research advice.