Comment Text:
CFTC Comment Submission
March 16, 2026
The Honorable Michael Selig Chairman
Commodity Futures Trading Commission
Three Lafayette Centre
1155 21st Street, NW
Washington, DC 20581
Re: Advance Notice of Proposed Rulemaking, RIN 3038-AF65
Dear Chairman Selig:
I. Introduction
I respectfully submit this comment urging the Commodity Futures Trading Commission ("CFTC" or "Commission") to adopt rulemaking requiring designated contract markets ("DCMs") with prediction markets to publish trade and market data in machine-readable formats, and to maintain and make available historical data archives accessible to the public.
As a researcher of prediction markets, I have spent considerable time studying the data publication practices of DCMs offering event contracts. What I have found is troubling: some exchanges publish daily trade data exclusively in PDF format, while others provide only a rolling single-day window of data, with no mechanism for the public to access historical records. These practices are incompatible with the Commission’s longstanding commitment to market transparency, public accountability, and informed regulatory oversight.
This comment identifies the problem, explains why it matters, and proposes practical, low-burden remedies.
II. Background and Current State of Data Publication Practices
The CFTC requires DCMs to maintain transparency as a core obligation under the Commodity Exchange Act ("CEA"). Core Principle 8 obligates exchanges to make relevant information available to market participants and the public. However, the specific format, frequency, and retention duration of publicly available data are not uniformly specified, and certain DCMs have interpreted their obligations narrowly.
Based on my direct review of currently operating prediction market DCMs, the following deficiencies are prevalent:
*PDF-only publication: Some exchange publishes its daily contract-level trade data exclusively in PDF format. PDFs are designed for human reading, not data processing. Fields embedded in a PDF table cannot be reliably extracted, sorted, queried, or aggregated without significant manual effort or error-prone parsing workarounds.
*Single-day rolling windows: Some exchanges publish only the current day’s data at any given time, with no persistent archive. Once the calendar rolls over, prior-day data becomes inaccessible to the public through official channels. There is no downloadable archive, no API endpoint for historical queries, and no FTP or bulk-access mechanism.
These practices create a two-tiered information environment: sophisticated actors with direct market access or proprietary data arrangements can reconstruct histories, while the general public cannot.
III. Why Machine-Readable and Historical Data Are in the Public Interest
Prediction markets now operate on questions of significant public importance: election outcomes, federal policy decisions, economic indicators, public health events, and more. The CFTC has authorized DCMs to offer event contracts on these topics precisely because of their potential informational value to the public. But that value is only realizable if the underlying data is accessible.
A. Academic and Policy Research
The academic literature on prediction market accuracy depends on access to granular, historical, machine-readable data. Accuracy calculations, calibration analysis, and cross-platform comparisons are unnecessarily difficult when data is locked in PDFs or purged after 24 hours. Without accessible historical data, it cannot be independently evaluated whether prediction markets are performing the price discovery function that justifies their authorization.
B. Regulatory Oversight and Market Integrity
Surveillance for market manipulation, wash trading, and anomalous volume patterns requires access to time-series data. A trader, whistleblower, journalist, or academic acting in the public interest cannot identify suspicious trading patterns if the data window is limited to a single calendar day. Historical machine-readable data enables the kind of retrospective analysis that is essential to market integrity enforcement.
C. Democratic Accountability
Event contracts on elections, legislation, and regulatory outcomes are of inherent public interest. Citizens, journalists, and advocacy organizations have a legitimate stake in understanding how these markets behave over time. Restricting data to a one-day window and burying it in PDFs effectively limits meaningful scrutiny to those with either deep technical resources or privileged market access. This is inconsistent with the Commission’s public interest mandate.
IV. Proposed Requirements
I urge the Commission to adopt the following requirements through rulemaking under Core Principles 2 (Compliance with Rules), 8 (Monitoring of Trading and Trade Processing), and 12 (Protection of Markets and Market Participants) of the CEA, or through a separate transparency rulemaking:
1. Machine-Readable Format Requirement: All DCMs offering prediction market contracts must publish daily trade data — including contract name, trade price, trade volume, open interest, and settlement price — in at least one machine-readable format such as CSV, JSON, or a structured text file with a defined schema. PDF-only publication does not satisfy this requirement.
2. Historical Data Retention and Accessibility: DCMs must retain and make publicly accessible at least 24 months of daily trade data in machine-readable format, updated no less frequently than once per trading day. Exchanges with existing historical data going back further should make that data available as well. Data must be hosted in a manner that permits bulk download without requiring API keys, account registration, or fees.
3. Standardized Schema: The Commission should publish or adopt a standard schema for prediction market daily trade data, including at minimum: contract identifier, underlying event description, contract expiration date, daily open/high/low/close prices, daily volume, open interest, and settlement price or outcome. Standardization across exchanges enables cross-platform research and public surveillance.
4. Versioning: Corrections to previously published data should be documented in a revision log.
V. Response to Potential Objections
"These requirements impose an undue burden on small exchanges." The technical burden of publishing a CSV file alongside a PDF is negligible. CSV generation is a basic function of any database or data management system. Any exchange sophisticated enough to operate as a CFTC-regulated DCM is capable of exporting data in structured format. The incremental cost is minimal and clearly outweighed by the public benefit.
"Historical data retention raises storage concerns." Two years of daily prediction market data in CSV format, even for a large exchange, amounts to merely megabytes of storage. Cloud hosting costs for this data are trivial. This objection does not withstand scrutiny.
"Market participants can purchase data from vendors." Requiring the public to pay for access to data generated by CFTC-licensed exchanges operating on matters of significant public interest is inconsistent with the Commission’s transparency obligations. Vendor dependency also introduces latency, quality concerns, and access barriers that fall disproportionately on non-commercial users.
VI. Conclusion
The prediction market industry has grown substantially and now trades on questions of central public concern. The data generated by these markets belongs in the public domain in a form that the public can actually use. Requiring machine-readable format and historical data availability is a modest, proportionate, and long overdue transparency measure.
I respectfully urge the Commission to adopt rulemaking requiring all DCMs offering event contracts to publish data in machine-readable formats and maintain publicly accessible historical archives of at least 24 months.
Respectfully submitted,
Mick Bransfield
Kingston Chase Consulting, LLC
