Coverage
What you get
| Item | Detail |
|---|---|
| Product code | STK-SURVIVORSHIP |
| Coverage | Delisted US stocks, plus membership history for major indices such as SPX, NDX and RUT |
| Granularity | One record per listing or per index membership spell; delisted prices as daily and 1-minute bars |
| History | Full membership and delisting history in the dataset; delisted symbols in bar endpoints from Pro |
| Delivery | REST, plus a one-off dataset (CSV and Parquet) |
| Update speed | Daily, as listings and index changes happen |
Run a momentum strategy on today's S&P 500 members back to 2010 and it will look great. It's also fiction. You've only tested the companies that survived and grew big enough to be in the index today, and quietly skipped every name that collapsed along the way. This dataset puts those names back.
Schema
Schema
Each record is one spell: a stock's listed life, or its time inside an index. A stock that joined and left the S&P 500 twice has two membership records.
| Field | Type | Required | Description |
|---|---|---|---|
symbol |
string | yes | Ticker as last traded |
index_code |
string | optional | e.g. SPX, NDX, RUT (membership records) |
start_date |
string | yes | Date listed or added to index |
end_date |
string | optional | Date delisted or removed; null if current |
delist_reason |
string | optional | merger, bankruptcy, exchange_delisting, going_private |
successor_symbol |
string | optional | Acquirer or new ticker |
permanent_id |
string | yes | Our ID that never changes when tickers are reused |
name |
string | yes | Company name |
Example record:
{
"symbol": "SIVB",
"index_code": "SPX",
"start_date": "2018-03-19",
"end_date": "2023-03-15",
"delist_reason": "bankruptcy",
"successor_symbol": null,
"permanent_id": "MD000184213",
"name": "SVB Financial Group"
}
permanent_id matters more than it looks. Tickers get reused: a symbol that belonged to a bankrupt bank can belong to a different company years later. Join your price history on permanent_id, not symbol, and the two never get mixed.
Sample data
Sample data
No free sample of this dataset is posted yet. Index membership history may need a licence from the index owner, so we're holding the file until that's settled. A sample file is available on request.
Use cases
Use cases
Point-in-time index backtests. For each rebalance date, select records where index_code = SPX, start_date is on or before that date and end_date is null or after it. That's the real universe on that day, failures included.
Modelling what happens at delisting. delist_reason separates bankruptcy from merger and going_private. A position in a bankrupt stock goes to near zero; one in an acquired stock gets the deal price. Treating both the same skews results.
Following a ticker through a merger. When successor_symbol is filled, you know where the shares went. Chain the records and you can carry a holding through an acquisition instead of letting it vanish from the portfolio.
Measuring the bias itself. Run your strategy twice, once on survivors only and once on the full set. The gap between the two equity curves is the number you would have been fooled by.
Code
Code examples
Example: S&P 500 members on a given date.
curl "https://api.marketdata.co/v1/stocks/index-membership?index_code=SPX&as_of=2020-03-02" \
-H "Authorization: Bearer $MARKETDATA_API_KEY"
Example: pull every delisted stock and count them by reason.
import os
import requests
import pandas as pd
resp = requests.get(
"https://api.marketdata.co/v1/stocks/delisted",
params={"start": "2000-01-01", "end": "2026-11-18", "limit": 10000},
headers={"Authorization": f"Bearer {os.environ['MARKETDATA_API_KEY']}"},
)
resp.raise_for_status()
gone = pd.DataFrame(resp.json()["results"])
gone["end_date"] = pd.to_datetime(gone["end_date"])
print(gone["delist_reason"].value_counts())
print(gone.groupby(gone["end_date"].dt.year)["permanent_id"].count().tail(10))
Pricing
Pricing
| Plan | Survivorship-bias-free data |
|---|---|
| Pro, $79/mo | Delisted symbols appear in our bar endpoints, plus index membership history |
| Business, $249/mo | Same, for commercial internal use |
Prefer a file? The planned one-off dataset is $299: delisted US stocks with daily and 1-minute bars plus index membership history, yours to keep. This is an early access price and may change before launch.
For context, the few providers we studied with a published price charge $29.99 to $59.95 a month, with a median of $52.50, based on only 3 prices. Two give a partial list away free.
Side by side
How we compare
Only 8 of the 50 providers we studied sell delisted stocks or index membership history, and some deliver it through a desktop app or an annual subscription. We plan to sell it as a one-off file you keep, with the price on this page, keyed by a permanent_id that also appears in our daily and intraday bars. Same field names, same symbols, no annual lock-in.
FAQ
Frequently asked questions
What is survivorship bias in stock data?
It's the error you get when a backtest only includes companies that still exist. Failed and acquired companies are missing, so returns look higher and drawdowns look smaller than they really were. Our survivorship bias guide walks through an example.
Are delisted stock prices split-adjusted?
Yes, bars for delisted names follow the same rules as live ones: split-adjusted OHLCV with raw prices alongside. See split and dividend adjustments for the details.
Which indices have membership history?
The schema covers index codes such as SPX, NDX and RUT. Which indices ship at launch depends on licences from the index owners, and we'll list them on this page when they're confirmed.
What time zone are the dates in?
start_date and end_date are plain trading dates (YYYY-MM-DD), so there's nothing to convert. Intraday bars for delisted names use UTC timestamps like the rest of our data.
Is this real-time data?
No. Delistings and index changes are dated events, and records are updated daily as they happen. Licensing here is about index membership, not speed.
What formats is the dataset in?
CSV and Parquet. The API returns JSON with the same field names.
Keep reading
Related
- Asset hub: Stocks
- Symbols: AAPL, MSFT, NVDA, AMZN, GOOGL, META
- Docs: API reference: stock bars
- Learn: Survivorship bias in backtests, How to backtest with 1-minute data