Finance or Gambling? Prediction Markets After Loper Bright and the Administrative Record

After Loper Bright displaced Chevron deference, how an agency constructs and defends the administrative record became a central question of administrative law, and prediction markets have become one of its most contested settings. As Americans increasingly trade on outcomes ranging from sports and entertainment to elections and geopolitical events, regulators and courts struggle over whether these products should be governed as financial instruments or as gambling. Prediction markets now facilitate substantial trading volume, command significant valuations, and operate amid active litigation across multiple states and federal courts. Regulators continue to assess these products through the lenses of hedging, price discovery, and information aggregation, while critics characterize them as gambling. Reframing that classification dispute as a problem of administrative law, this Article examines how it manifests not in theory, but in the administrative record itself. Recent empirical work using large language model (LLM)–based computational text analysis demonstrates that regulatory outcomes track the structure, complexity, and informational content of public comments, and that certain forms of participation are more likely to shape the administrative record and, in turn, agency reasoning. By treating the notice-and-comment record as a legally consequential evidentiary object rather than a passive aggregation of preferences, this Article demonstrates how linguistic framing shapes the factual predicates on which agencies must now defend regulatory choices under independent judicial judgment. The empirical inquiry applies computational text and sentiment analysis to public comments submitted on the Commodity Futures Trading Commission’s (CFTC) 2024 Event Contracts Notice of Proposed Rulemaking. Although the CFTC withdrew the proposal on February 4, 2026, the resulting notice-and-comment record remains a durable evidentiary source shaping litigation and future governance. Using natural language processing, LLM classification, multi-model validation, and confidence-weighted semantic analysis, the study assesses whether commenters characterize prediction markets primarily as economic instruments or emphasize their gambling-like features and risks. The analysis reveals a pronounced asymmetry: of 812 submissions, 87.2% addressed economic, financial, or regulatory consequences rather than gambling-related characteristics, employing terminology central to financial discourse. Only 5.8% emphasized gambling risks or entertainment dynamics, with high-confidence gambling terminology, with the remaining 7% mixed or neutral. This economic-dominant framing appears across all stakeholder positions (opponents of the rule, supporters of the restriction, and neutral commenters), suggesting that regulatory disagreement reflects not competing conceptualizations of what event contracts are, but rather how a shared economic understanding should be regulated. The dataset also reveals skewed participation: while the median comment length was 147 words, the average exceeded 2,500 words, indicating that most participants submitted brief statements while a small minority provided extensive filings. One hundred eighty comments included attachments, some representing coordinated filings by elected officials or organizations. Professional industry representatives and organized stakeholders, though a numerical minority, dominate the record through the length and density of their submissions, amplifying the financial framing presented to regulators. Despite substantial scholarship documenting gambling-like mechanisms, the administrative record reflects predominantly financial framing. These findings illuminate how stakeholder framing shapes the factual record upon which classification decisions rest. Because courts now exercise independent judgment over both statutory meaning and the adequacy of agency fact-finding, In a post–Loper Bright environment, where courts apply heightened scrutiny to agency fact-finding and statutory interpretation, the administrative record's rhetorical construction can prove decisive. By demonstrating empirically how language and framing constitute the evidentiary foundation for regulatory classification, this Article reveals that the legal status of prediction markets may turn less on their inherent characteristics than on the discursive strategies through which interested parties present them to regulators and courts, a dynamic that reaches any agency defending a contested classification on the record. This dynamic raises foundational questions about whether notice-and-comment produces records courts can treat as genuinely independent evidentiary foundations.