Historical CFTC Crypto Enforcement Actions

Study CFTC crypto enforcement by separating the action announced now, allegations, settlement state, and stated monetary relief at the source-document level.

Historical CFTC crypto enforcement actions can be isolated from a normalized cross-agency feed without treating every press release as the same legal outcome. Alphanume's Crypto Enforcement dataset covers CFTC enforcement releases from 2024 forward alongside SEC and DOJ sources. Filter agency=cftc, then preserve action type, allegation categories, respondent rows, settlement state, relief basis, and the agency document URL.

The Crypto Enforcement documentation defines the action and respondent keys. It also states the essential legal guardrail: a complaint or charge records allegations, while a plea, verdict, or settled order records a different stage. The data do not justify calling a party liable before the public action does.

Separate actions from respondents

The CFTC release is keyed by action_key. Each named respondent receives a separate record_id and respondent index. A release naming several parties therefore produces several rows. Use distinct action keys for action counts and record IDs for party counts.

Bans, registration restrictions, cooperation adjustments, and nonmonetary remedies may appear in source prose but are not separate served fields in the current contract. If those concepts are required, create reviewed research columns from document_url and retain the quoted evidence. Do not infer them from action type or relief amount.

Retrieve the CFTC slice

Date filters apply to published_date, the structural date of the agency release. The extracted action_date has its own precision enum and may describe an underlying legal event. Use publication date for public-event timing unless the research explicitly studies the legal-process date.

GET /v1/crypto/enforcement
agency=cftc
date_gte=2024-01-01
date_lte=2025-12-31

action_table = deduplicate by action_key
respondent_table = retain every record_id

group action_table by action_type
report allegation categories separately
retain settled_flag, monetary_relief_usd,
       monetary_relief_basis, document_url

Pagination uses published date, action key, and respondent index. Return all three cursor values. This keeps multi-respondent actions contiguous and avoids losing the final respondent on one page or repeating it on the next.

Compare stages without inventing a resolution field

For each action, treat action_type as the normalized stage. A settled order can be grouped separately from another action type, while settled_flag preserves what the document says. If a charge later produces a related resolution, follow-up fields can link successive actions within the same agency and respondent key. They do not perform cross-agency case matching.

• Report charges, settlements, pleas, and verdicts as distinct stages.
• Keep unknown settlement states separate from stated false.
• Review source documents for bans or cooperation treatment.
• Retain all allegation categories rather than only the primary one.

respondent_key is normalized text, not resolved legal identity. Review potential collisions and name changes before creating a party history. Asset names are also document literals, so a separate mapping is needed to combine spelling or ticker variants.

Follow-up analysis should retain both ends of each served chain. Record the prior action key, next action key, and days to follow-up without assuming that the chain is a complete case docket. The linkage follows the same agency and normalized respondent name. It can miss name changes and does not join a parallel DOJ or SEC matter. A reviewed cross-agency case map belongs in a separate table with explicit evidence.

Prevent monetary and legal overstatement

The numerical failure mode is summing relief across respondent rows. One action's amount repeats on every named party, so aggregate distinct action-key and amount pairs. Relief can combine restitution, disgorgement, penalties, forfeiture, or interest. Keep the basis beside the number and avoid comparing totals as if every dollar represented the same sanction.

The legal failure mode is describing allegations as judgments. Use stage-specific language and link to the agency document. Coverage starts in 2024, which means the output is a defined recent history rather than the CFTC's full digital-asset enforcement record.

Publication counts should not be read as a direct measure of regulatory severity. One complex action and several narrow actions affect the count differently, while press-release practices can change over time. Report counts, respondent breadth, allegation mix, relief basis, and stage as separate dimensions. If a cooperation adjustment is manually coded, publish the decision rule and source excerpt so another researcher can reproduce the classification.

Build an auditable CFTC action ledger

As a concrete next action, retrieve a completed year and produce one action-level table plus one respondent-level table. Count actions by stage, cross-tab allegation categories, and report settlement unknowns. Review every source document in a small sample for any separately coded ban or cooperation field. Reconcile relief after action-level de-duplication. The ledger can then support comparisons without merging charges, settlements, and remedies into one unsupported label.