Why tickers are missing from so many filings — and how we fix them

Pull a raw Periodic Transaction Report and you will notice something inconvenient: there is no ticker field. Members disclose descriptions of assets, not symbols. Sometimes the filing system captures a ticker; very often it does not. If you have ever wondered why free scrapes of this data are full of blank symbols, this is why.

Where the gaps come from

• Free text is free. The same company arrives as “NVIDIA Corporation - Common Stock,” “NVIDIA CORP,” or “Nvidia” — filed by different offices, typed by different staffers.
• Not everything has a ticker. Municipal bonds, private placements, real-estate partnerships, and cash-equivalent funds are all disclosable, and many simply do not map to an exchange symbol.
• Corporate actions rot the mapping. Mergers, spin-offs, renames, and delistings mean the right symbol for a 2015 filing is not always the right symbol today.
• Some filings are pictures. A share of Senate paper filings surface as scanned documents, so before you can match an asset, you have to read it.

How our resolution pipeline works

We treat ticker resolution as a first-class data problem rather than a regex afterthought:

1. Normalize. Strip the boilerplate (“- Common Stock,” “Class A,” punctuation soup) and canonicalize the issuer name.
2. Match against an issuer database that carries name histories, so “Facebook Inc” and “Meta Platforms” resolve to the same issuer at the right point in time.
3. Score and gate. High-confidence matches publish automatically. Ambiguous ones — and there is always an ambiguous tail — queue for human review before a symbol is attached.
4. Never discard the original. Every trade keeps the asset description exactly as filed, next to the ticker we resolved. You can always check our work against the record.

Why it matters

Symbols are how the data becomes usable: they are what lets you ask “who in Congress traded NVDA this quarter?” or wire a watchlist alert to your portfolio. A dataset that silently drops the hard 30% of filings is not a smaller dataset — it is a biased one, because the messy filings are not randomly distributed across members and asset classes. Resolving the tail is tedious. It is also the product.