Summary

  • The Commodity Futures Trading Commission’s enforcement actions fell nearly 80% in 2025 from the prior decade’s annual average, and the agency approved prediction markets at triple the historical rate in the same period, while its workforce stood 21% below the 10-year average by year-end 2025.
  • Then-acting Chair Caroline Pham told enforcement staff the agency was “closing all our crypto [cases],” according to 26-year enforcement lawyer Joe Konizeski, and the agency subsequently dismissed Biden-era actions against Gemini, Celsius, FTX, and WorldWideMarkets.
  • Three hypotheses fit the documented pattern — capacity collapse from the workforce drop, directed policy reorientation toward the cryptocurrency industry, or both operating simultaneously — and the categorical closure directive plus the named-case dismissals tilt the working reading toward the combined explanation.
  • Senator Elizabeth Warren requested a Government Accountability Office investigation of CFTC staffing in July, the GAO plans to “initiate” the probe in December, and Chairman Michael Selig has moved the agency toward unilateral rulemaking on digital assets after the Senate failed to advance a crypto and banking regulation bill.

By every indicator NPR’s reporting assembles, the CFTC’s enforcement footprint contracted sharply while its approval queue for prediction markets expanded at a comparable rate, and the same period saw a named categorical directive from agency leadership and the targeted departure of the attorneys who worked the cases that directive named. The Commodity Futures Trading Commission’s enforcement actions fell nearly 80% in 2025 from the prior decade’s annual average per NPR’s accounting, with the agency’s workforce at 21% below its 10-year average by year-end 2025 according to U.S. Office of Personnel Management data; six prediction markets were approved in 2025, three times the prior decade’s average annual rate, and six more have been authorized in 2026 with 18 applications pending. Five published indicators — workforce data, action counts, directorial statements, case-level dismissals, and an Inspector General finding — converge on the central question the Government Accountability Office intends to examine when it initiates a probe in December: whether the documented pattern reflects capacity collapse from understaffing, directed policy reorientation, or both mechanisms operating in the same window.

Three competing explanations for the enforcement decline

The substrate supports three readings.

The first is capacity exhaustion. Under the documented headcount, the agency could not staff the investigations those dropped actions represent, and the dismissals of Biden-era actions against Gemini, Celsius, FTX, and the WorldWideMarkets foreign-exchange matter are cases that could not be pursued to settlement or trial. The workforce figure, the action-count figure, and the Inspector General finding all fit a single capacity mechanism; the prediction-market approvals are a separate queue processable regardless of enforcement posture because the two operations depend on different staffing pools. CFTC Inspector General Christopher Skinner strengthened the workforce premise from inside the agency’s oversight apparatus, writing in the agency’s recent annual report that “The CFTC should consider human capital management as a top priority” and that the agency needs “the appropriate talent and skill sets to meet an expanding regulatory landscape.”

The second is directional reorientation. NPR’s reporting carries Jeff Le Rich, a CFTC enforcement lawyer between 2005 and 2025, on the record: “The Trump administration had run partially on a platform that it would be friendlier to the crypto industry”; “in order to show results to that industry, some people were punished for bringing those cases”; “The end result was that a lot of the people that worked on those cases either left or were forced out”; “It castrated the enforcement team”; “Many of the people who were targeted were high-performing attorneys who had brought some of the most consequential and complicated cases over the past few years.” Konizeski, who NPR reports was one of the dozens forced out in 2025, supplied the on-record specification: “And so, of course, everybody who had a crypto scam had to close it and any matter that was even marginally related to crypto ended up getting closed.” Pham left the CFTC in December 2025 and joined the cryptocurrency company MoonPay as chief legal and administrative officer; she did not respond to NPR’s request for comment on whether she gave preferential treatment to the crypto industry.

The third is both mechanisms operating in the same window. A pure capacity-exhaustion reading does not derive categorical directives from a staffing model; under that reading, case closures would distribute across the agency’s portfolio rather than concentrate in the category named by leadership. With the categorical directive and the case-cluster dismissals both present, the combined hypothesis fits the substrate more parsimoniously than either explanation alone. If the Pham statement were removed, the dismissals of named Biden-era actions would still support a non-capacity reading, but the inferential leap from observation to motive would lengthen; the directive sharpens the combined reading because the workforce shrink, the enforcement-collapse figure, the named-case-category directive, and the named-case dismissals all hold simultaneously. A cross-reading from inside the agency’s oversight apparatus — Skinner’s IG finding on human-capital priorities — sits parallel to, rather than in tension with, the directional reading because headcount strain is itself a finding about workforce management, not a finding about enforcement choices.

A directorial signal set sharpens the directional reading without resolving it. Selig declared at a White House event in the Roosevelt Room in August, flanked by crypto executives, “The era of political lawfare, debanking and regulation by enforcement is over”; “Innovators, like the people in this room, are welcome to the White House — not railroaded to the big house”; and that the United States needs to “dominate” these new financial technologies and “write the rules that define the next generation of financial markets.” Selig is the sole sitting commissioner of a normally five-member body, giving him unchallenged power over the agency’s direction. Before his time at the CFTC, he was a corporate lawyer for cryptocurrency firms and prediction markets; his financial disclosure report shows he sold between $65,000 and $150,000 in Bitcoin before he became a commissioner — a public-record act that reduces any inference of concealment on that transaction while leaving the directional reading otherwise intact.

Selig’s rulemaking commitment — that “the American people deserve clarity, certainty, and consumer protections in the crypto asset markets” and that the agency is committed to “delivering this by incorporating crypto asset transactions into its uniform national market regulatory framework” — sits in direct tension with the on-record capacity ceiling described by former CFTC director Brian Young, who said the headcount cut “limits the agency’s ability to write new rules” and that “At the end of the day, they need people to process these applications.” Either the rulemaking is materially under-resourced relative to its stated scope — in which case it leans more heavily on industry consultation and existing precedent, compounding the directional-shift concern — or it issues in a form the agency itself cannot fully staff to enforce. Either reading sharpens the burden the GAO probe carries.

Cascade of consequences across time horizons

In the immediate horizon, the CFTC approved six prediction markets in 2025, three times the prior decade’s average, and authorized six more in 2026, with 18 applications pending. As of July 2026, the workforce remained 16% below its decade average, or about 100 staffers short; the agency’s stated plan is to hire roughly 100 employees in mission-critical areas by the end of 2026, while the current job list comprises seven full-time positions. CFTC spokesman Zach Fulton told NPR the agency “is on track to hire roughly 100 employees in mission critical areas by the end of 2026” and that “a number of CFTC staff opted to step away from public service in recent years.” Per-application review depth declines as the queue grows; conditions attached to approved platforms have less capacity to address advertising-conduct questions of the kind Konizeski identified when he accused Polymarket and Kalshi of “using fraudulent means” to target young men and asserted that “CFTC knows this and allows it to continue” — neither platform responded to NPR’s requests for comment. A separate testimonial stream reinforces the consensus read without resolving which hypothesis dominates: a former CFTC employee who speaks regularly with current staff and spoke to NPR on the condition of anonymity reported the agency is unable under current headcount to execute President Trump’s 2025 executive order on cryptocurrency — “My friends that are there are dying” — a strain on the executive-order directive that is itself a directional finding, not a generic capacity claim.

In the one-to-three-year horizon, the GAO’s December probe terminates either with a resource-only finding or with findings extending into directional conduct; the latter path shortens the window before Congress uses its appropriations and subpoena authorities. Senator Warren, on the record, framed the question as whether “the watchdog is actually out there doing a little barking, or has the watchdog been locked up somewhere back in the barn?” and said Congress holds the “power to investigate” but that “led by the Republicans, Congress not only doesn’t have a backbone, it doesn’t even have any interest in doing it.” Asked whether a future Democratic Congress would subpoena the heads of the CFTC, the crypto companies, and the prediction markets, Warren said, “it is our job to use all those tools to investigate.”

In the three-plus-year horizon, three plausible branches appear. Under a reinforcing branch, the institutional legitimacy of the CFTC as the federal derivatives overseer contracts, with spillover into the commodity-futures domain that predated crypto. Under a second reinforcing branch, foreign regulators — the U.K. Financial Conduct Authority and EU bodies acting under ESMA — extend jurisdictional reach into U.S.-touching activity as the U.S. supervisory regime contracts, though the substrate does not document such moves and that gap is named explicitly. Under a dampening branch, the CFTC rebuilds headcount and resumes prior enforcement categories with political cover from GAO findings, leaving directional-shift concerns to be resolved through future commission composition. The productive cross-horizon conflict is visible: short-term approval velocity trades against long-term market-integrity risk, and the documentary record does not show which side the agency is weighing.

Pressure points the substrate does not resolve

Several assertions remain allegations. The platform-to-personnel causal chain — Le Rich’s claim that the reduction “was not a coincidence” because “the Trump administration had run partially on a platform that it would be friendlier to the crypto industry” — depends on transition-team communications, OPM records on separations, and contemporaneous internal CFTC emails that the substrate does not contain; the December probe’s scope as NPR describes it does not yet extend to that chain. A GAO investigator would need such documentation to substantiate the timing inference.

Case-level allegations also remain allegations. Konizeski’s claim that Polymarket and Kalshi used “fraudulent means” to target young men is reported with the platforms’ non-response rather than independently established conduct. The Pham-after-MoonPay timing sequence — directing agency case closures while serving as acting chair, then joining a cryptocurrency company in a senior legal role shortly thereafter — is the factual configuration NPR flagged when it asked whether Pham gave preferential treatment to the crypto industry, and Pham did not respond. That question does not fit inside the staffing-focused probe NPR describes for the GAO, and the substrate does not report any open Office of Government Ethics or CFTC Inspector General review of the move; the dimension may be addressed along a procedural channel that does not depend on the December timeline.

The article foregrounds crypto and prediction-market enforcement but underweights the derivative-market context the CFTC has regulated historically. Konizeski’s “Fewer cops on the beat… creates a real incentive to start engaging in fraudulent and deceitful behavior” framing is offered in support of a prediction-market point, and the article explicitly notes the CFTC “largely focused on more traditional derivative markets, such as grain and stock futures contracts” before the recent expansion. The implication for traditional commodity-futures enforcement under the headcount drop is not separately addressed.

The single-commissioner governance structure — Selig as the sole sitting member of a normally five-member body — is symmetrical in its reading: it enables institutional direction-setting by a single officeholder as readily as it enables reform, and the GAO investigation is positioned to test the structure’s operation in practice. The CFTC’s stated plan to make rules without congressional input, following the Senate’s failure to advance a major crypto and banking regulation bill over the summer, compresses the deliberative surface area at the same moment the deliberative depth of the agency itself contracts — sole commissioner, shrinking workforce, eighteen pending prediction-market applications, six new approvals already authorized in 2026.

Five developments that would materially shift the posterior

Five follow-on developments would update the working reading: the GAO investigation’s documentary findings, with initiation scheduled for December 2026; the disposition of any congressional subpoenas and oversight activity; the agency’s progress toward its stated 100-employee hiring goal by year-end 2026; any additional pre-commission financial disclosures by Selig beyond the existing record; and the documented incidence of fraud in prediction and crypto markets during 2026. The substrate does not resolve any of these; the December GAO probe is positioned to resolve the first.

The diagnostic test of which hypothesis the substrate supports is fairly mechanical. If GAO’s first interim communication narrows the probe to documentation of attrition rather than to directives on case closures, the resource branch dominates the immediate future; if it extends to those directives, the directional branch dominates regardless of staff additions. The first signal will arrive in GAO’s first interim communication.

Analytical techniques used in this piece

This analysis applies the methods below. Each links to a short, plain-English explainer you can read and reuse.

Bayesian Hypothesis Network
Updates the probabilities of competing hypotheses as evidence accumulates.
Consequences & Sequels
Plays a decision forward to its first- and second-order consequences.
Red-Team Assessment
Models a capable adversary probing a plan for the seams they would exploit.
Bayesian Reasoning
Starting from base rates and updating beliefs proportionally as evidence arrives.