OTC volume: separate reporting venues from ownership
An OTC trading row identifies a reporting unit or venue-related activity, not the investor who ultimately owns the shares. One operator can be connected with several rows, while a non-attributed bucket can contain several firms. A useful reconstruction separates reporting identity, execution role and ownership before adding volume.
This guide uses FINRA’s OTC Transparency page and ATS MPID FAQ as read on October 25, 2026. The weekly numbers and operators below are fictional. It concerns FINRA’s described equity trade-reporting dataset, not a complete record of investor holdings or every form of OTC financial activity.
Identify the reporting object
FINRA’s OTC Transparency description states that information is derived from OTC trades reported to its equity reporting facilities and published on a delayed basis for ATSs and relevant member firms. Certain de minimis non-ATS activity is aggregated without attribution.
Record security, reporting period, ATS or non-ATS category, MPID or other provided reporting identifier, legal operator and whether the row is a component or total. An issuer CIK and a trade-reporting MPID perform different identification jobs.
The ATS MPID FAQ, questions 1–3 explains the unique identifier requirement and its specific debt-versus-equity exception. Two identifiers are not automatically two ATSs or two independent beneficial investors.
Preserve the source’s roles rather than treating every appearance of a firm name as an execution venue. The same legal organization may operate a venue, route orders, act as a counterparty or hold a different reporting role.
Work a nonoverlapping volume total
Assume the following published-week exercise for one security. All figures are shares, the periods and coverage match, and the component rows are mutually exclusive under the stated data definitions.
| Row | Shares | Role in calculation |
|---|---|---|
| ATS Alpha | 300 | Component |
| ATS Beta | 200 | Component |
| Non-ATS activity | 500 | Component total |
| OTC total | 1,000 | Sum of the three components |
| Compatible all-market total | 2,500 | Separate scale denominator |
OTC share of the matched all-market activity = (300 + 200 + 500) / 2,500 = 40%.
Adding the displayed OTC total of 1,000 to its own components produces 2,000 and double counts the same activity. The total row is a check, not an additional venue.
The 40% is valid only under the explicit compatibility assumptions. The all-market denominator must use the same security, period, units and relevant trade-counting conventions. A monthly consolidated total, a different share class or unmatched corrections cannot be substituted for the assumed 2,500.
The result is a trading-activity share. It does not establish that an institution bought 1,000 shares, increased holdings by 1,000 or owns 40% of the issuer.
Separate an operator map from an owner map
Suppose independent identity evidence establishes that Operator X operates both Alpha and Beta. Their combined venue activity is 500, or 20% of the assumed all-market volume. That is activity associated with X’s two venues under this map, not a 500-share investment held by X.
Customers can trade on a venue without making the operator the beneficial owner. A share can also trade repeatedly during the week. Volume does not identify the distinct shares held at the end of the period.
If the legal operators differ despite a common brand, keep them separate until a supported grouping is appropriate for the research question. A brand-level grouping can be useful but must not erase the underlying legal identities or be labeled an investor position.
A movement from one venue to another could change reported rows without revealing a change in final investors’ aggregate holdings. The venue data alone does not connect the two sides to a common beneficiary.
Treat non-attributed volume as a bucket
Assume the non-ATS total of 500 comprises named activity of 450 and a non-attributed bucket of 50. The 50 is already inside the 500; adding it again would produce an incorrect OTC total of 1,050.
The bucket is not a single anonymous institution. It may combine multiple eligible firms, and the public aggregate does not supply their individual contributions or customers. A user cannot assign the entire 50 to the largest named firm.
The absence of a firm’s separate row can therefore reflect aggregation rather than zero activity. Before treating a missing name as inactivity, inspect the dataset’s applicable publication and attribution rules.
This distinction also limits concentration calculations. A ranking of attributed rows describes the visible reporting units. It does not fully rank every underlying firm or beneficial owner hidden by the aggregation.
A routed order changes the role interpretation
The FAQ’s third example discusses an ATS routing orders outside itself and appearing only as the contra-side executing broker in the reported trade under specified circumstances. It explains when an additional MPID is unnecessary.
Consequently, an ATS identifier appearing as a counterparty does not, by itself, establish that the execution occurred within that ATS. Read the field role and reporting context before assigning a venue.
For a fictional order with a portion matched internally and another portion routed away, keep the executions and reporting roles separate. Do not count the whole order as internal matching merely because an ATS initiated the routing. Nor should a routed component be added twice under the sender and receiver without examining the reporting basis.
This is a role boundary, not evidence that the published official totals themselves necessarily double count. The error can arise in an analyst’s reconstruction when unlike fields are treated as compatible volume rows.
Keep counts and average sizes on their proper denominators
Assume Alpha’s 300 shares represent six trades, Beta’s 200 represent four and the non-ATS 500 represent twenty. Average sizes are 50, 50 and 25 shares respectively.
The combined average is total shares 1,000 divided by thirty trades, about 33.33 shares. The simple average of the three row averages is about 41.67 and weights the rows rather than the trades.
Neither average reveals the number of distinct investors. One investor can generate multiple trades, and a trade can involve intermediaries. A column headed “trades” is not an owner count or a position count.
FINRA also provides quarterly ATS statistics. Do not mix a quarterly total with a selected week merely because the venue identifier agrees. The period is part of the data key.
Execute the reporting-unit worksheet
| Field | Retain | Inference controlled |
|---|---|---|
| Security and period | Exact class, dates and units | Matched numerator and denominator |
| Reporting unit | ATS/non-ATS identifier and legal entity | Row not called beneficial owner |
| Role | Venue, reporting side or contra-side | Routed activity not mislabeled internal matching |
| Hierarchy | Component, bucket and total | No total-plus-component addition |
| Attribution | Named or non-attributed | Missing name not automatically zero |
| Availability | Published period and retrieval date | Delayed data not treated as real time |
| Comparison | Same scope and revisions | No unmatched market share |
Public data can establish the reported activity and identities made available under its rules. Order-level routing histories and beneficial ownership links usually require different evidence. Do not bridge that gap by matching familiar names or similar amounts.
When totals disagree, compare period endpoints, attribution, security identity, component hierarchy and revisions. Document whether the selected denominator includes all the relevant activity on a compatible basis.
Update the map when an operator changes, a reporting identifier is clarified, aggregation rules or source files change, or a correction appears. Preserve old identities and versions so a change of reporting unit is not silently presented as investor accumulation or liquidation.
Test the hierarchy before calculating concentration
Create a parent identifier for the non-ATS total and mark its attributed and non-attributed children. A summation should include the parent or the children, never both. Then keep a separate legal-operator map with an effective date, so changes of ownership or identifiers do not rewrite prior weeks silently.
For concentration, state whether the unit is a venue, legal operator or attributed reporting firm. Combining Alpha and Beta under Operator X answers a different question from ranking the two venues separately. Neither supplies the concentration of beneficial investor holdings, even if every reporting unit is successfully mapped.
FAQ
Is a venue’s volume its operator’s stockholding?
No. Reporting activity does not establish beneficial ownership.
Can a non-attributed bucket be treated as one firm?
No. Its underlying firms and contributions are not identified by that aggregate.