betaFeedback

WorksComputer Science

Can Language Models Identify Shadow Trading Targets? An NLP Evaluation of SEC Enforcement Theory

Abstract

Shadow trading – trading in a peer firm's securities on the basis of material nonpublic information (MNPI) about an "economically linked" company – is a novel and contested theory of insider trading liability, first prosecuted in SEC v. Panuwat (2023). Enforcing it requires identifying economically linked firms ex ante, a determination the SEC makes only after the fact using mass market surveillance infrastructure. We ask whether NLP can do what the SEC's theory presumes insiders already know: identify peer firms ex ante from publicly mandated disclosures. Using a two-stage LLM pipeline applied to Item 7 (Management's Discussion and Analysis) sections of SEC 10-K filings, we score semantic similarity across 30 M&A events spanning five industries and relate similarity to announcement-day abnormal stock returns. On the Panuwat fact pattern itself the pipeline recovers Incyte among the closest peers, a sanity check on the one case with a known outcome. Across the full dataset, however, we find no association: pooling 217 peer observations, the within-event rank correlation between similarity and abnormal return is +0.07 (permutation p = 0.37), and the mean per-event Spearman correlation is +0.05 with a 95% confidence interval of [-0.08, +0.18] – narrow enough to exclude any moderate relationship rather than merely failing to detect one. A case-level reading agrees: 14 of 30 events support the hypothesis, 12 contradict it, and 4 are ambiguous. We also find that Incyte fell outside the standard $2B-$10B mid-cap band on the day before the announcement, complicating the "mid-cap oncology" category the SEC invoked. These results are exploratory and bound to this pipeline, corpus, and return measure, but they put pressure on the empirical premise of shadow trading enforcement and bear on constitutional questions surrounding the SEC's financial surveillance infrastructure.

AI4Law @ ICML 2026, Seoul, South Korea. Copyright 2026 by the author(s).

Introduction

On August 18, 2016, Matthew Panuwat, a business-development executive at Medivation, Inc., received an email from his CEO indicating that Pfizer would soon likely acquire the company. Seven minutes later, he purchased call options—not in Medivation, but in Incyte Corporation, a separate publicly traded biopharmaceutical company with no direct business relationship to Medivation. When the acquisition was announced the next day, Incyte’s stock rose 7.7%, and Panuwat sold his options for approximately $107,000 in profit ([30]).

The U.S. Securities and Exchange Commission (SEC) charged Panuwat with insider trading under a new theory it calls “shadow trading”: that MNPI about Firm A is material to any firm with which A is “economically linked,” such that trading in Firm B’s securities on the basis of such information constitutes securities fraud ([30]).

The enforcement asymmetry. Classical insider trading enforcement begins with a known event: regulators detect abnormal volume in a stock, then subpoena the relevant broker’s records to identify the trader. Shadow trading inverts this logic. The suspicious trade may occur in any security whose connection to the MNPI event is asserted only after the fact. Regulators therefore cannot index on a single security; they must surveil trading activity across the entire market to find which trades correlate with any given corporate event. Among the SEC’s enforcement programs, this is unique: flash-crash reconstruction starts from a known event, classical and misappropriation insider trading start from a tagged trade in a particular security, but shadow trading is the only use case for which the SEC must search the corpus before it can identify which trader to investigate. That is precisely what the Consolidated Audit Trail (CAT) enables: a comprehensive, queryable record of every order, cancellation, and execution in U.S. equities and options markets, covering more than 100 million investors [17]. Multiple legal challenges have characterized CAT as a “dangerous dog” [18] and a potential Fourth Amendment violation [17, 26]; this paper’s empirical results bear on those challenges directly.

The NLP question. If economic linkage can be identified ex ante from public disclosures, targeted watchlists could replace mass surveillance. Because the SEC requires firms to disclose their risk-reward profiles in annual reports, Item 7 (Management’s Discussion and Analysis) of the Form 10-K is an ideal corpus for testing this hypothesis: if two firms genuinely occupy the same strategic niche, their MD&A sections should exhibit high semantic similarity, and if the shadow trading hypothesis is empirically grounded, those same firms should experience positive abnormal returns when an acquisition of one is announced. We operationalize this test using a two-stage LLM pipeline: a peer discovery stage that identifies candidate companies from filings in the SEC’s EDGAR database, followed by a document-grounded similarity ranking stage that assigns semantic similarity scores. We then compute announcement-day abnormal returns for each peer firm and assess whether textual similarity predicts peer stock reactions across 30 acquisition events in five industries.

Contributions.

  1. A fully documented two-stage LLM pipeline for long-context multi-document financial similarity analysis, including principled handling of Foreign Private Issuers (FPIs). The prompts, rubric, and per-event outputs are reproduced in full in the Appendix; because the underlying model is closed-weight, exact score reproduction is not guaranteed (Section 7).

  1. A dataset of 30 M&A events (2013–2023) with LLM-derived similarity scores and manually computed abnormal returns, spanning biopharmaceuticals, consumer staples, technology, finance, and automotive industries.

  1. Empirical evidence that, for this pipeline and corpus, textual similarity in SEC filings does not predict announcement-day peer stock reactions: a pooled within-event rank correlation of +0.07 (p=0.37p = 0.37) and a mean per-event Spearman correlation of +0.05, 95% CI [−0.08, +0.18], with 14/30 events supportive under a case-level reading. Cross-industry variation is pronounced—Technology yields zero supporting cases across five acquisitions while Automotives yields zero contradictions—though with five events per sector we treat this as descriptive rather than established.

  1. A correction to the record on the Panuwat fact pattern: Incyte’s capitalization on the day before the announcement was $14.3B, outside the standard 2B−2B-10B mid-cap band that the SEC’s “mid-cap oncology” framing presupposes.

This work addresses a high-stakes legal enforcement question where the answer—whether language can serve as a legal standard—has direct constitutional consequences for the largest government-mandated collection of personal financial data in U.S. history.

Background

Shadow Trading and SEC v. Panuwat

Under the classical theory of insider trading, liability attaches to a relationship between the trader and the issuer of the traded security. United States v. O’Hagan [24] extended this to misappropriation: a person who trades on MNPI obtained from a source owed a duty of trust and confidence commits securities fraud even if they have no relationship with the issuer. SEC v. Panuwat [30] pushes this further still: the duty-source need not be connected to the security traded at all. Because Panuwat owed a duty to Medivation, and because the court found that Medivation and Incyte were “economically linked” mid-cap oncology companies, his MNPI about Medivation was deemed material to Incyte’s securities. Table 1 shows the case timeline.

DateEvent
Aug. 2016Panuwat trades Incyte options, 7 minutes after CEO email
Aug. 2021SEC files civil complaint
Jan. 2022Motion to dismiss denied
Nov. 2023Motion for summary judgment denied
Apr. 2024Jury finds Panuwat liable after 8-day trial
Sep. 2024Civil penalty of $321,197 imposed
Nov. 2024Panuwat appeals (9th Cir. No. 24-6882)
May 2025U.S. Chamber of Commerce files amicus brief
Jun. 2026Ninth Circuit hears oral argument
Aug. 2026Panel decision pending

Table 1. Timeline of SEC v. Panuwat.

The key doctrinal problem is that neither firm mentioned the other in its annual report, and neither shared a direct business relationship. “Economic linkage” was established retroactively by the SEC’s expert witness, who testified that Incyte’s 7.7% announcement-day return “was not caused by normal fluctuations.” The category itself—“mid-cap oncology”—is not a recognized industry classification, and the record fixes no capitalization threshold for it. The opinion describes Medivation as a mid-cap, oncology-focused biopharmaceutical company and returns repeatedly to the scarcity of mid-cap oncology assets, but nowhere states what capitalization range the term denotes [30]. The label does substantive work—it is what makes Incyte comparable enough for MNPI about Medivation to be material to Incyte’s securities—while remaining numerically undefined. The economic stakes are large: [3] document positive abnormal options volume in approximately 25% of U.S. takeovers between 1996 and 2012, and [16] estimate single-event shadow-trading profits ranging from $139,400 to $678,000. Since Panuwat, the only concluded enforcement action has been Bechtolsheim, which settled for a $923,740 civil penalty [31]; even so, roughly one in five public companies has now formally prohibited shadow trading in its insider-trading policy [12].

The Consolidated Audit Trail and Fair Notice

Because shadow trading liability may attach to any “economically linked” security, the theory requires whole-market retrospective surveillance rather than targeted investigation. The Consolidated Audit Trail (CAT)—created in 2012 and covering every equity and options order across all U.S. securities exchanges [20]—is the infrastructure that makes this possible. Shadow trading is unique among the SEC’s enforcement use cases in requiring this inversion; other programs proceed from a known event or a flagged trade in a particular security. Shadow trading thus transforms CAT from a regulatory tool into what critics call a “digital general warrant,” and raises a direct fair notice problem: for a regulatory prohibition to withstand due process scrutiny, a person of ordinary intelligence must be able to determine what conduct is prohibited before the fact [27]. If the set of prohibited securities cannot be identified from public disclosures ex ante, an insider has no principled way to know which trades are off-limits. This paper tests that empirical premise directly.

Text-Based Firm Similarity in Financial Disclosure

The SEC’s “economic linkage” standard presupposes that similarity between firms is readable from their public disclosures. The foundational evidence for this comes from the Hoberg-Phillips Text-based Network Industry Classification (TNIC) [10], which demonstrates that pairwise firm similarity constructed from mandatory 10-K product descriptions captures competitive dynamics far better than static SIC codes. If TNIC-style analysis can classify industries, it should, in principle, identify the economically linked peers that shadow trading liability requires.

However, TNIC relies on bag-of-words similarity over Item 1 (Business Description), which is semantically blind to synonymous terminology: “oncology” and “cancer treatment” would be treated as unrelated tokens. This creates the administrability gap—if two firms describe the same strategic niche using different technical vocabularies, BoW similarity fails to detect the linkage that shadow trading liability presupposes.

Learned models have been applied to paired financial-report comparison before: [14] train encoders to compare 10-K pairs for forecasting, establishing that document-scale report comparison is not itself novel. What long-context generative LLMs add is the ability to ingest complete Item 7 sections without chunking and to bridge vocabulary differences explicitly, applied here to the forward-looking MD&A narrative rather than the backward-looking Item 1. Our contribution is not the comparison architecture but the use of it to operationalize a legal standard and test that standard against market outcomes. Whether closing the administrability gap is sufficient to make economic linkage reliably identifiable ex ante is the empirical question this paper addresses.

Methodology

Dataset Construction

We assembled 30 major public acquisitions across five industries: biopharmaceuticals (10 events), consumer staples (5), technology (5), finance (5), and automotive (5), spanning 2013–2023. Biopharmaceutical events are the primary test set, as shadow trading doctrine originated there; the remaining industries serve as a control group. Candidate events were generated by prompting Gemini 3.1 Pro for acquisitions structurally comparable to Pfizer/Medivation in target size and deal structure. From that pool, retained events met four criteria: the target was a U.S.-listed public company with an EDGAR filing history; the deal was an announced acquisition of the whole company rather than a divestiture or minority stake; the announcement date could be fixed unambiguously; and the sector had enough comparable public peers to populate a candidate list. This universe was not pre-registered and was not drawn by a mechanical screen from a deal database, so selection effects cannot be ruled out. Additionally, the candidate pool reflects which transactions are most prominent in the model’s training data, which compounds the contamination concern discussed in Section 7. Events were fixed before abnormal returns were computed. For each acquisition target (Firm A), we retrieved the most recent Form 10-K filed on EDGAR prior to the announcement date, using Item 7 (MD&A) and Item 7A as the primary similarity corpus. For Foreign Private Issuers (FPIs) filing Form 20-F, we substitute with Item 5 (“Operating and Financial Review and Prospects”) and Item 11 (“Quantitative and Qualitative Disclosures About Market Risks”) accordingly.

Two-Stage LLM Pipeline

Figure 1 illustrates our Augmented Hoberg-Phillips (HP) framework, which extends the classical HP bag-of-words approach [10] by substituting a long-context LLM for cosine similarity over binary term vectors, and shifting from Item 1 (Business Description) to Item 7 (MD&A) to better reflect the forward-looking narrative an insider would possess. We decompose the task into two sequential stages to avoid self-consistency bias: models anchored to prior numerical scores resist revision even when new evidence warrants it [11, 9, 13]. Both stages are implemented in Google AI Studio using Gemini 3.1 Pro in isolated per-event sessions; full model configuration details are provided in Section 3.3.

Augmented Hoberg-Phillips Framework

Figure 1. Augmented Hoberg-Phillips Framework

Stage 1 — Peer Discovery. The target firm’s Item 7 is uploaded to identify ten peer companies as of the acquisition date. Google Search grounding is enabled to confirm that candidate firms were publicly traded on a U.S. exchange and had an active filing obligation at the time, and specified in the prompt. Section 7 discusses how this bounds the “identifiable from public filings” framing. The model provides natural language dimensional justifications (Table 2) without numerical scores to mitigate anchoring bias. Key constraints include: no hindsight (market state at announcement only); reasonable insider standard; acquirer excluded; active 10-K/20-F filing obligation; and NYSE/NASDAQ trading status required. Full prompts are in Appendix A.

DimensionFocusAssessment Criteria
Tech OverlapProductSimilarity of platforms or pipelines.
StrategicFinancialMatching growth stages and revenue models.
RiskExternalGeographic exposure and headwinds.

Table 2. Stage 1 Dimensional Justification Rubric

Stage 2 — Document-Grounded Ranking. Item 7 sections for all identified peers are uploaded as a single multi-document prompt. The model assigns scores in [0,1][0,1] based

exclusively on the provided disclosures—mirroring the closed-corpus logic of HP [10]—with Google Search disabled. Outputs are parsed into structured .csv tables. We weight Technical/Product Overlap most heavily (40%) to reflect the SEC’s emphasis on shared therapeutic mechanisms in Panuwat; Financial (Strategic Stage) and Risk & Market (Exposure) each receive 30%, capturing the forward-looking narrative dimensions unique to Item 7 relative to Item 1.

DimensionWeightAssessment Criteria
Technical40%Core technology and product engines.
Financial30%Growth trajectories and margin profiles.
Risk/Market30%Geographic focus and specific headwinds.

Table 3. Stage 2 Similarity Score Rubric

Model and Configuration

We select Gemini 3.1 Pro [7] for its 1M-token context window (sufficient to ingest multiple complete 10-K sections without chunking), robust multi-document cross-referencing, and state-of-the-art reasoning performance (77.1% on ARC-AGI-2; 94.3% on GPQA Diamond). These capabilities matter because traditional bag-of-words approaches are semantically blind, creating the administrability gap our study seeks to measure [5]. The pipeline runs in Google AI Studio with Thinking Level set to High, enabling Chain-of-Thought reasoning over dense SEC disclosures [34]. Each acquisition event runs in an isolated session to prevent cross-contamination. Top-P is slightly lowered from 0.95 to 0.90 to reduce uncertain token probability mass while preserving domain-specific financial vocabulary [15]; all other parameters remain at default.

Abnormal Return Calculation

For each peer firm, we compute the announcement-day abnormal return:

ARi=Ri(0)−Rbenchmark(0)(1)AR_i = R_i^{(0)} - R_{\mathrm{benchmark}}^{(0)} \tag*{(1)}

where Ri(0)=(Pi(0)−Pi(−1))/Pi(−1)R_i^{(0)} = (P_i^{(0)} - P_i^{(-1)})/P_i^{(-1)} is the peer firm’s raw return on announcement day (Day 0) relative to the prior trading day (Day −1-1), and Rbenchmark(0)R_{\mathrm{benchmark}}^{(0)} is the same-day return of a sector-level ETF: XBI (biopharmaceuticals), IYK (consumer staples), XLK (technology), XLF (finance), and IYC (automotive). This directly mirrors the SEC economist’s methodology at the Panuwat trial. We compute the Spearman rank correlation (ρ\rho) between the similarity score and ARiAR_i per event (Section 4.5) and generate industry-specific scatter plots (Appendices C and D).

Mid-cap validation. We computed the market capitalization of every identified peer as of the day before its acquisition announcement, using contemporaneous share counts wherever they were retrievable.2 Only 30.9% of peers fall within the generic $2B–$10B mid-cap band, and 30.9% within S&P’s eligibility range for additions to the MidCap 400 as it stood at the time of the Panuwat trade, approximately $1.4B–$5.9B (S&P Dow Jones Indices, 2026). The two bands select overlapping but not identical sets of firms; that they capture the same share is coincidental, and the point is simply that the conclusion does not turn on which mainstream definition of “mid-cap” one adopts. Within the biopharmaceutical subset the shares are lower still, at 21.7% and 24.6% respectively.

This undercuts the “mid-cap” label as an ex ante constraint in two ways. First, the band captures a minority of the peer universe, so it does little to narrow which firms an insider would know to trade. Second, and more directly, Incyte itself falls outside both definitions: on the day before the Medivation announcement its capitalization was $14.3B—above the $2B–$10B band, and more than twice the $5.9B ceiling S&P then applied to MidCap 400 additions. “Mid-cap oncology” is not a recognized industry classification, and under neither mid-cap convention in force in 2016 does it contain the security actually traded. The two capitalization figures that do appear in the record point the same way. An investment bankers’ chart grouped Medivation and Incyte among a small set of commercial oncology companies spanning roughly $5B to $75B in market capitalization—a range running from mid-cap through mega-cap—and the opinion elsewhere places the pair in the $10 billion range while describing them as mid-cap assets (United States District Court, N.D. Cal., 2023). A $10B-range firm sits at the very ceiling of the generic band and well above S&P’s 2016 threshold; neither figure in the record is consistent with mid-cap as the term was conventionally used at the time. Because the record fixes no threshold of its own (Section 2.1), the classification also turns on which vintage of an external definition one applies: S&P revises its eligibility ranges quarterly, and by 2026 the MidCap 400 addition band had risen to $8B–$22.7B—a range that would admit a $14.3B firm. Whether a 2016 trade fell inside the category must be assessed against the 2016 bands; a threshold drawn from a later period cannot establish what an insider could have known at the time. Whatever work the label does in characterizing the link between Medivation and Incyte, it is not the work of putting an insider on notice ex ante.

Reproducibility. All prompts, per-event CSV outputs, and scatter plots are provided in the Appendix. Abnormal return calculations are implemented in Python using yfinance and pandas, with Stooq as a fallback for delisted tickers. Because Gemini 3.1 Pro is a closed-weight model, exact score reproduction is not guaranteed across API versions; we therefore treat LLM outputs as structured qualitative judgments and compute all financial metrics independently.

Results

Reference Case: Pfizer/Medivation (2016)

We treat the Pfizer/Medivation acquisition—the factual basis of SEC v. Panuwat—as a reference case, the one event in our dataset with a known legal outcome against which to sanity-check the pipeline. Figure 2 shows the scatter plot of similarity rank versus abnormal return. The pipeline identifies Tesaro (TSRO, Rank 1), Exelixis (EXEL, Rank 2), and Incyte (INCY, Rank 3) as the most similar peers; all three show positive abnormal returns (+1.60%, +1.33%, and +5.04% respectively). Incyte—the stock Panuwat actually traded—is ranked third and posts the largest abnormal return among the top three ranked peers. It is not the largest in the event overall: Puma Biotechnology (PBYI), ranked ninth, returned +6.44%.

Scatter plot of abnormal return by similarity rank for Pfizer/Medivation (2016)

Figure 2. Pfizer/Medivation (2016): all three top-ranked peers (TSRO, EXEL, INCY) posted positive abnormal returns, which is what makes this case read as supportive under the procedure of Section 4.2. The fitted trendline is negative but essentially flat—it falls 0.36 percentage points across the full rank range (R2=0.002R^2 = 0.002)—and the rank correlation over the same peers is slightly negative (ρ=−0.10\rho= -0.10), driven by the ninth-ranked outlier PBYI. Neither statistic supports the hypothesis on its own.

Three caveats bound what this case can establish. Incyte is recovered but is not top-ranked; the per-event rank correlation is in fact slightly negative (ρ=−0.10\rho= -0.10, Section 4.5) because of the PBYI outlier; and SEC v. Panuwat has been widely reported since 2021, so a model with parametric knowledge of the case may associate Medivation with Incyte for reasons unrelated to the filing text we supply. We therefore read this case as a sanity check that the pipeline surfaces plausible oncology peers, not as validation of the methodology. @bibliography_keys_used united_states_district_court_2023_sec_v_panuwat | s_p_2026_s_p_u_s

Full Results: 30 Acquisition Events

Table 4 summarizes all 30 events. Our interpretation of whether a case supports or contradicts the SEC’s theory is not a mechanical byproduct of the OLS trendline alone. We applied the following decision procedure to each event: (i) compute the OLS trendline of abnormal return on similarity rank; (ii) inspect the signs of the abnormal returns for the three highest-ranked peers; (iii) where (i) and (ii) disagree, or where the trendline is driven by a single peer, weigh the absolute scale of the returns and the share of peers moving in each direction. An event is labeled supports when the trend and the top-ranked reactions both point in the direction the shadow trading hypothesis predicts, contradicts when they both point against it, and mixed when they conflict and no reading dominates.

IndustryAcquisition AnnouncementYearPeersTrendTop-3 AR DirectionInterpretation
BiopharmaceuticalsPfizer / Medivation†20165↓\downarrow+ + +Supports (reference)
BiopharmaceuticalsAbbVie / ImmunoGen20238↓\downarrow+ + +Strongly supports
BiopharmaceuticalsBMS / Karuna20238↓\downarrow+ − ≈ 0Supports
BiopharmaceuticalsBMS / Mirati20237↓\downarrow+ − +Contradicts
BiopharmaceuticalsBMS / MyoKardia20208↓\downarrow+ ≈ 0 −Strongly contradicts
BiopharmaceuticalsBMS / Turning Point20229↓\downarrow+ + +Supports
BiopharmaceuticalsGilead / Forty Seven20207↑\uparrow− + ≈ 0Contradicts
BiopharmaceuticalsMerck / Acceleron20217↓\downarrow≈ 0 − +Strongly contradicts
BiopharmaceuticalsMerck / Pandion20217↓\downarrow+ + +Strongly supports
BiopharmaceuticalsPfizer / Seagen20237↑\uparrow≈ 0 − −Contradicts
Consumer StaplesAmazon / Whole Foods20179↓\downarrow− − ≈ 0Supports
Consumer StaplesDanone / White Wave Foods20166↓\downarrow+ ≈ 0 ≈ 0Supports
Consumer StaplesJ.M. Smucker / Hostess20235↓\downarrow+ + +Supports
Consumer StaplesKroger / Albertsons20229↓\downarrow+ − +Contradicts
Consumer StaplesReckitt / Mead-Johnson20179↑\uparrow≈ 0 − −Contradicts
TechnologyAMD / Xilinx20209↑\uparrow+ − −Contradicts
TechnologyAvago / Broadcom20155↑\uparrow− − +Contradicts
TechnologyCisco / Splunk202310↓\downarrow− + +Mixed
TechnologyMicrochip / Microsemi20186↑\uparrow− − −Mixed
TechnologySS&C Technologies / DST Systems20187↑\uparrow+ ≈ 0 ≈ 0Contradicts
AutomotivesApollo / Tenneco20229↑\uparrow− + −Mixed
AutomotivesGoodyear / Cooper20219↓\downarrow+ + +Supports
AutomotivesTenneco / FDML20187↓\downarrow− + −Supports
AutomotivesWarner / Delphi20207↑\uparrow− ≈ 0 +Mixed
AutomotivesWarner / Remy20158↓\downarrow+ + ≈ 0Strongly supports
FinanceBB&T / SunTrust20197↑\uparrow+ + +Contradicts
FinanceKeyCorp / FirstNiagara20155↓\downarrow+ − −Strongly supports
FinancePeople / United20196↑\uparrow− ≈ 0 ≈ 0Contradicts
FinanceSterling / Astoria20175↓\downarrow+ − −Supports
FinanceWSFS / Beneficial20186↓\downarrow+ + +Supports
SummarySupports: 14Contradicts: 12 Mixed: 4

Table 4. Summary of all 30 acquisition events. Trend: ↓\downarrow = negative (similarity predicts positive AR); ↑\uparrow = positive (similarity predicts negative AR). †Factual basis of SEC v. Panuwat; used as a reference case (Section 4.1). Per-event rank correlations for these same events are reported in Table 12.

We state plainly that step (iii) involves judgment. The labels were assigned by the authors jointly, after inspecting the return data rather than under a rule fixed in advance, so they carry researcher degrees of freedom that a pre-registered rule would not. This is precisely why we do not rest the paper’s quantitative claim on them. Section 4.5 reports a fully mechanical rank-correlation statistic computed from the same per-event data, and readers who distrust the labels can read the headline result from that analysis alone; the two agree on the bottom line.

A “negative” trendline (↓\downarrow) indicates that higher similarity scores correlate with higher abnormal returns—the core requirement for the shadow trading theory. Conversely, a “positive” trendline (↑\uparrow) indicates that higher similarity scores correlate with lower abnormal returns. However, if a trend is negative but the absolute returns for the top-ranked peers are negligible or negative, the case does not “support” a theory of predictable profit. Peer counts per event vary below ten for several reasons: firms identified in Stage 1 were subsequently excluded when their EDGAR filings lacked sufficient Item 7 content, or when historical price data was unavailable due to post-announcement delisting.

Biopharmaceutical Results

Our analysis begins with the Biopharmaceutical sector, the very crucible of the SEC’s shadow trading doctrine. In the Pfizer/Medivation acquisition, the pipeline identified Incyte (the firm traded by Panuwat) as a highly similar peer that experienced a positive abnormal return (Figure 2). This shows that the similarity metric can surface a firm the SEC regarded as economically linked, subject to the caveats in Section 4.1.

However, when the same methodology is applied across ten major biopharmaceutical acquisitions, that relationship does not hold. As summarized in Table 4, the results are close to a coin flip: exactly half of the transactions (five out of ten) contradict the shadow trading hypothesis, and the mean rank correlation for the sector is +0.12 (Table 6), not distinguishable from zero. In these contradictory cases, the firms with the highest 10-K semantic similarity to the target either experienced negligible price movement or negative abnormal returns on the announcement day.

IndustryEventsMean ρ\rhoρ>0\rho> 0
Biopharmaceuticals10+0.117+0.1176/10
Consumer Staples5+0.156+0.1564/5
Technology5−0.239-0.2391/5
Automotives5+0.085+0.0853/5
Finance5+0.036+0.0363/5
Total30+0.04617/30

Table 6. Mean per-event Spearman correlation by industry. ρ>0\rho> 0 is the direction the shadow trading hypothesis predicts.

For example, while transactions like AbbVie’s acquisition of ImmunoGen generated the uniform sympathy rally the SEC expects (strong positive returns for the top three peers), deals like Pfizer’s acquisition of Seagen and Gilead’s acquisition of Forty Seven yielded fundamentally unaligned market reactions. In the Pfizer/Seagen transaction, the top three most semantically similar peers registered zero or negative abnormal returns. This directly refutes the premise that shared fundamental business traits—as disclosed in mandatory SEC filings—reliably dictate correlated market behavior.

These findings carry direct implications for NLP-based enforcement heuristics. As reported in Section 3.4, only 21.7% of the biopharmaceutical peers our pipeline identifies satisfy the standard 2B2\mathrm{B}–$10B\$10\mathrm{B} mid-cap definition, so the cohort is not structurally homogeneous in the way the Panuwat framing suggests. Even so, a state-of-the-art LLM pipeline operating on the SEC’s own mandated disclosures predicts peer stock reactions at chance level within this sector—suggesting that “economic linkage,” as operationalized through disclosure text by this pipeline, is an unreliable signal rather than a consistent market phenomenon.

Cross-Industry Results

To determine whether the shadow trading phenomenon extends beyond the Biopharmaceutical sector, we expanded our NLP analysis to twenty additional acquisitions across four diverse industries: Consumer Staples, Technology, Automotives, and Finance. As detailed in Table 4, the expansion of the dataset further erodes the structural validity of the SEC’s “economic linkage” theory, with results distributed across supporting, contradicting, and mixed outcomes.

The breakdown by sector reveals profound inconsistencies that undermine any notion of a universal, predictable market mechanism. For example, deals like AMD/Xilinx and Avago/Broadcom produced inverted trend lines where highly similar peers saw negligible or negative returns. Similarly, the Consumer Staples and Finance sectors exhibited near-random distributions of support and contradiction. In the BB&T/SunTrust banking merger, a robust sympathy rally lifted the entire cohort, yet the magnitude of those gains completely decoupled from semantic similarity, yielding a strongly contradictory trend line. As shown in Table 5, Technology is the only sector with zero supporting cases while Automotives yields zero contradictions. One reading is that document-recoverable linkage is industry-contingent, with tight supplier ecosystems in automotive manufacturing producing spillovers that dispersed platform-based technology firms do not. We flag this as a hypothesis rather than a finding: each sector contributes five events, and the corresponding rank correlations (Table 6) are not distinguishable from zero.

IndustryEventsSupportsContradictsMixed
Biopharmaceuticals10550
Consumer Staples5320
Technology5032
Automotives5302
Finance5320
Total3014124

Table 5. Results by industry.

Across the full 30-event dataset, semantic similarity predicts peer stock movement direction in 14/30 cases (46.7%), counting the four mixed events as non-supporting. A two-sided binomial test against H0=0.5H_0 = 0.5 yields p≈0.86p \approx0.86; excluding the mixed events instead gives 14/26 and p≈0.85p \approx0.85, so nothing turns on that convention. We stress that this is a failure to reject rather than a demonstration of no effect: with n=30n = 30, a binomial test of this form has little power against modest departures from chance, and absence of evidence is not evidence of absence. The rank-correlation analysis in Section 4.5 provides the sharper statement, bounding how large an association the data can accommodate. Taken together, the results suggest that “economic linkage” as our pipeline operationalizes it is not a stable, document-recoverable feature of firm relationships but an idiosyncratic outcome that varies across industries, deal structures, and market conditions.

Rank-Correlation Analysis

The labels in Table 4 involve judgment (Section 4.2). To give the central claim a form that does not, we compute for each event the Spearman rank correlation ρ\rho between the LLM similarity score and the announcement-day abnormal return, over the same per-event data reported in Appendices C and D. Throughout this section we adopt the convention that ρ>0\rho> 0 means more similar peers earned higher abnormal returns—the direction the shadow trading hypothesis predicts. This is the opposite sign from the trendline arrows in Table 4, which regress return on similarity rank: because rank 1 denotes the most similar peer, rank runs opposite to score, and the same underlying relationship therefore appears with the opposite slope.

Significance is assessed by a permutation test. Holding the similarity scores fixed, we reshuffle the abnormal returns among the peers of the same event and recompute ρ\rho; the pp-value is the fraction of reshuffles yielding a correlation at least as extreme as the one observed. This assumes only that, under the null, any return could equally well have attached to any peer—no distributional assumption is required. For events with n≤8n \le8 peers we enumerate all n!n! reassignments exactly; for the eight events with nine or ten peers we sample 200,000 at random. Per-event values appear in Appendix B.

Pooled result. Events differ in peer count, so before pooling we convert each event’s similarity scores and abnormal returns to within-event ranks and standardize them, placing every event on a common scale. The correlation across all 217 standardized observations is ρpooled=+0.066\rho_{\mathrm{pooled}} = +0.066 (permutation p=0.37p = 0.37, reshuffling returns within event). Taking instead the 30 per-event correlations as the unit of analysis, their mean is ρˉ=+0.046\bar{\rho} = +0.046 with standard deviation 0.348; the corresponding 95% confidence interval on that mean,

formed as ρˉ±t29,0.975 s/30\bar{\rho} \pm t_{29, 0.975}\,s/\sqrt{30}, is [−0.084,+0.175][-0.084, +0.175]. Seventeen of thirty events have ρ>0\rho> 0, twelve have ρ<0\rho< 0, and one is exactly zero (two-sided sign test p=0.58p = 0.58). Only one event—AbbVie/ImmunoGen (ρ=+0.81\rho= +0.81, p=0.022p = 0.022)—is individually significant at α=0.05\alpha= 0.05, against the 1.5 expected by chance alone across thirty tests.

What this rules out. The confidence interval is the substantive result. It is centered near zero and narrow enough to exclude any association stronger than ρ≈0.18\rho\approx0.18. This is a stronger statement than that null results support: rather than merely failing to detect a relationship, the data are inconsistent with a moderate or large one. An enforcement standard resting on the premise that filing similarity identifies the firms that will move on an announcement requires an association far stronger than the upper end of this interval.

Agreement with the qualitative labels. Of the 25 events assigned a decisive label and a nonzero ρ\rho, the sign of ρ\rho agrees with the label in 19 (76%). The six disagreements are informative rather than alarming: they arise where the OLS trendline and the rank correlation weigh outliers differently. Pfizer/Medivation is one such case—the trendline is negative, but ρ=−0.10\rho= -0.10, because Puma Biotechnology (rank 9) posted the single largest abnormal return in the event. That the mechanical statistic and the holistic labels disagree on individual events while converging on the same aggregate conclusion is the appropriate reading: the aggregate null is robust to how one scores any particular case.

The industry pattern in Table 6 is directionally consistent with the label-based breakdown—Technology is the only sector with a negative mean—but the per-industry samples are five events each, and none of these means is distinguishable from zero. We report them as description, not as evidence of sector-specific effects.

Discussion

LLM similarity as an enforcement heuristic. By construction, our pipeline can bridge synonymous terminology that bag-of-words similarity treats as unrelated. In cases like AbbVie/ImmunoGen, textual similarity tracks announcement-day abnormal returns closely. But across the corpus the association is indistinguishable from zero (ρpooled=+0.07\rho_{\mathrm{pooled}} = +0.07, p=0.37p = 0.37; 14/30 events supportive), and that is insufficient to serve as the basis for a legal standard: a standard right about as often as it is wrong provides neither the fair notice required by due process ([27]) nor the precompliance-review predicate that Patel requires for warrantless administrative access to business records ([25]). If some sector did exhibit a reliable signal, targeted watchlists would be preferable to whole-market collection on that ground alone—but our data do not identify such a sector. The per-industry correlations in Table 6 rest on five events each and none is distinguishable from zero, so we cannot recommend any sector as a candidate for a narrower list. A recent graph-based deep learning framework for shadow trading detection confirms the computational intensity of the problem ([23]), but scale alone does not resolve the constitutional defect—it accentuates it.

Failure mode analysis. Qualitative inspection of contradicting cases reveals two recurring failure modes. The first is sector-wide contagion: in BB&T/SunTrust, all peers rallied regardless of similarity rank because the merger signaled regulatory approval of large-bank consolidation generally, swamping any firm-specific signal. The second is outlier dominance: in BMS/MyoKardia, a single firm (CYTK, +14.3% AR) drove an apparent negative trendline while the remaining peers showed negligible returns, making the case technically “contradicting” by trendline but reflecting one idiosyncratic event rather than a systematic pattern. These failure modes suggest that even where 10-K similarity is high, macro-structural events and single-stock noise can overwhelm the announcement-day signal. Both are properties of the outcome measure rather than of the text, which is one reason we treat the one-day abnormal return as a limitation of the test rather than a verdict on the disclosures.

Why mass surveillance persists. The fundamental issue is not that similarity never tracks peer reactions—it does in some events—but that it does so unpredictably. An insider running our pipeline before trading would face genuine uncertainty about whether a given trade crosses the shadow trading line. The SEC, by contrast, can observe the announcement-day return first and construct a textual similarity argument post hoc—which is essentially the Panuwat fact pattern, where “mid-cap oncology” was an analyst category invoked retrospectively to explain Incyte’s price reaction. In January 2026 the SEC approved an amendment eliminating names, addresses, birth years, and taxpayer identifiers from CAT reporting and directing the deletion of customer data already collected ([32]). That change does not disturb the search-first/identify-later inversion: replacing directly identifying data with pseudonymous identifiers addresses the objection to holding personal information in bulk, but leaves untouched the structural defect that the trigger for re-identifying a trader cannot be specified ex ante from public disclosures.

Implications for live litigation. A state-of-the-art LLM, given the same mandated disclosures an insider would read, identifies shadow trading targets at a rate indistinguishable from chance in our sample. We are careful about what follows from this. Our result concerns one pipeline, one corpus, and one return measure; it does not establish that no method could recover economic linkage, nor that mass surveillance is legally required. What it does show is that the administrability premise—that a person of ordinary intelligence could determine ex ante which securities are off-limits—has not been demonstrated and is not supported by the most capable text-analytic tool we could bring to bear on the disclosures the doctrine itself points to. That is a question courts weighing Davidson v. Atkins ([19]) and the appellate posture of Panuwat will have to confront, and it is the contribution we claim.

Firm similarity from financial text. Prior work applying HP-style text similarity to financial disclosures has focused on market clustering, risk assessment, and portfolio optimization ([8]; [5]), and on evaluating LLMs applied to 10-K sections for question answering and information extraction tasks ([2]; [35]). Closest to our setup, [14] learn to compare pairs of financial reports for forecasting, showing that document-scale report comparison is an established task; their objective is predictive accuracy, whereas ours is whether a similarity ranking can discharge a legal identifiability requirement. None of this work evaluates whether textual firm similarity predicts announcement-day stock reactions, nor does it examine the legal implications of that relationship. We extend this lineage by applying long-context generative LLM reasoning to operationalize the SEC’s “economic linkage” standard and test it empirically against market outcomes.

Shadow trading in finance and law. [16] introduced “shadow trading” in the economics literature, estimating single-event profits of $139,400–$678,000; [3] documented positive abnormal options volume in ~25% of U.S. takeovers between 1996–2012. [29] discuss enforcement implications from recent SEC cases. [23] propose AM-GIN, a graph-based surveillance framework that models the market as a spatio-temporal graph integrating sectoral ties and price co-movements; unlike our approach, AMGIN operates on market structure data rather than disclosure text and is designed as a regulator surveillance tool rather than an ex ante identifiability test. To our knowledge, this paper is the first to apply LLMs to mandatory SEC disclosures to operationalize economic linkage as a legal standard and evaluate whether that textual signal predicts peer stock reactions across industries.

Privacy and surveillance in securities markets. [18] and [4] document privacy concerns with CAT; [17] challenges its Fourth Amendment constitutionality ([19]). [26] provides the Supreme Court framework for comprehensive digital surveillance, and in Chatrie v. United States ([28]) the Court extended it to an architecturally analogous bulk-then-unmask program, holding that querying a bulk repository for an individual’s location history is a Fourth Amendment search because a person retains a reasonable expectation of privacy in those records. CAT has the same architecture: identifying a single trader means querying a record assembled from everyone. [6] show that pseudonymization is a brittle defense—four spatio-temporal points identify 95% of individuals in a 1.5M-person corpus—a result that bears directly on CAT’s reliance on pseudonymous customer identifiers ([1, 21]).

Limitations

Several limitations bound the scope of our conclusions, and two of them are severe enough that we state them before the rest.

No baselines. We did not compare our pipeline against TNIC-style bag-of-words similarity ([10]), SIC/GICS industry peers, TF-IDF or embedding similarity, or random within-sector peer rankings. This is the most consequential gap in the study. Without such comparisons we cannot distinguish three explanations for our null result: that public filings do not encode the relevant linkage, that this particular LLM pipeline fails to extract it, or that announcement-day abnormal returns are too noisy a target for any similarity measure. Our claim concerns this pipeline, not NLP generally.

The similarity scores are not validated as measurements. Each event was scored once. We report no run-to-run variance, no sensitivity to prompt or rubric wording, no second model, and no agreement with human expert judgments. LLMs used as evaluators are known to exhibit systematic biases and position or verbosity effects ([33]), and our rubric weights (40/30/30) were chosen to reflect the Panuwat testimony rather than tuned or validated. Because

Gemini 3.1 Pro is closed-weight, its outputs are also not guaranteed stable across API versions. We therefore treat the scores as structured qualitative judgments rather than deterministic measurements, and a stability study is the first thing we would add.

Statistical power and the meaning of the null. A non-rejected null is not proof of no effect. Our binomial test on 14/30 (p≈0.86p \approx0.86) should not be read as demonstrating the absence of a relationship. The rank-correlation confidence interval in Section 4.5, [−0.08,+0.18-0.08, +0.18], is the defensible version of the claim: it bounds how large an association the data can accommodate. Industry-level statements rest on five events each and are descriptive only.

Construct validity of the outcome measure. We equate “economic linkage” with a same-day price response, which is narrower than the legal concept of materiality to another issuer’s securities. Linked firms might react over longer windows, through options markets, or through negative competitive channels. A perfect linkage measure could still fail to predict a one-day return dominated by liquidity and idiosyncratic noise; our negative result may therefore indict the proxy rather than identifiability. Relatedly, the expected sign is deal-dependent—in Amazon/Whole Foods we score a case as supportive when similar peers fell further, on the reasoning that the acquisition signaled competitive pressure—and we did not fix a sign convention by deal type in advance.

Peer discovery is not a clean filings-only condition. Stage 1 ran with Google Search grounding enabled to confirm historical listing status, and we did not log the retrieved sources. Filing-derived evidence, search-derived evidence, and the model’s parametric knowledge are therefore not fully separable, and for widely reported deals the model may recall the outcome. This bounds how strictly the “identifiable from public filings” framing should be read.

Sample and design. The dataset covers 30 events across five industries spanning 2013–2023 and omits sectors such as energy, real estate, and healthcare services. The event set was assembled by the authors rather than drawn from a pre-registered universe, and the supports/contradicts/mixed labels were assigned after inspecting returns (Section 4.2). Abnormal returns use sector ETFs rather than firm-specific market models—consistent with the SEC economist’s approach in Panuwat, but introducing benchmark risk for firms whose betas deviate from their sector. Finally, the market-capitalization figures in Section 3.4 exclude delisted firms and firms whose ticker symbols were later reassigned, which biases the surviving sample toward larger capitalizations.

Conclusion

We present an NLP-grounded empirical evaluation of the SEC’s shadow trading enforcement theory. On the Panuwat fact pattern our two-stage LLM pipeline recovers Incyte among the closest peers. Across 30 acquisition events, textual similarity in 10-K MD&A sections shows no association with announcement-day peer reactions: the pooled within-event rank correlation is +0.07 (p=0.37p = 0.37) and the mean per-event Spearman correlation is +0.05 with a 95% confidence interval of [−0.08, +0.18]. Under a case-level reading, 14 of 30 events support the hypothesis and 12 contradict it. The confidence interval is the operative result: it excludes any moderate association, rather than merely failing to find one.

We are deliberate about the scope of this claim. We tested one closed-weight model, one rubric, one filing section, and one abnormal-return measure, without baseline comparisons. Our evidence supports the conclusion that this pipeline does not recover economic linkage from Item 7 text in a way that predicts announcement-day returns. It does not establish that no NLP method could, and a null of this kind cannot prove the absence of an effect—only bound its plausible size.

Future work should add the baselines this study lacks (TNIC-style bag-of-words similarity, industry-code and embedding peers), test multiple models and repeated runs to quantify scoring stability, expand the dataset, and combine multiple 10-K sections (Items 1, 1A, and 7) or supplementary filings (Form 8-K, Schedule 13D). A promising inversion of our design would begin from the market rather than the filings: scan a broad universe for the largest abnormal movers around each announcement, then ask whether those firms share characteristics that were publicly knowable beforehand. That reverses the direction of inference and would test identifiability without presupposing that our peer-discovery stage found the right candidates.

The broader stake is unchanged by these caveats. Shadow trading liability presumes that an insider can determine in advance which securities are off-limits, and our results give no support to that premise: the most capable text-analytic tool we could bring to bear on the disclosures the doctrine itself points to recovers no usable signal, while a linkage argument remains easy to construct once the price reaction is known. That asymmetry—cheap to assert after the fact, unavailable before it—is what the evidence assembled here documents, and it is what a defense of the doctrine’s administrability now has to answer.

Impact Statement

This paper evaluates whether NLP can operationalize the SEC’s “economic linkage” standard for shadow trading enforcement—a legal theory with direct consequences for the largest government-mandated collection of personal financial data in U.S. history. We identify two categories of broader impact.

Beneficial impacts. Our findings provide courts, policymakers, and legal scholars with an empirical test of the predictability premise underlying shadow trading enforcement. If economic linkage cannot be reliably identified ex ante from public disclosures—as our results suggest—then targeted watchlists cannot replace the mass surveillance architecture of the Consolidated Audit Trail, and constitutional challenges to that architecture are empirically grounded. This has direct relevance to pending litigation (Davidson v. Atkins; Panuwat at the Ninth Circuit) and to ongoing debates about the scope of the SEC’s administrative authority. The dataset and two-stage pipeline documented in the Appendix may also benefit researchers studying NLP for financial regulation, legal AI, and text-based firm similarity.

Potential risks. A sophisticated actor could in principle adapt this approach to look for trading opportunities in sectors where the signal appears stronger. We note four mitigating factors: (1) all inputs—10-K filings—are already publicly available on EDGAR; (2) the overall association is indistinguishable from zero, and the apparent sector differences rest on five events each and are not statistically distinguishable from chance, so there is no validated high-signal sector to exploit; (3) any trading on the basis of publicly available disclosure text would not constitute insider trading under any current theory of liability; and (4) the pipeline is not a trading tool and should not be used as one—the results here are exploratory and provide no reliable basis for trading, portfolio construction, or enforcement targeting. We emphasize this last point because a negative result about administrability could be misread as a positive result about predictability in the cases that did line up; it is not. A secondary risk is that our findings could be read as endorsing shadow trading. We emphasize the contrary: our empirical result that the enforcement theory lacks a reliable ex ante basis raises constitutional questions about the infrastructure used to enforce it, not the underlying prohibition on trading on material nonpublic information.

Acknowledgments

We are grateful to Professor Sebastian Zimmeck (Wesleyan University) and Alex Abdo (Knight First Amendment Institute, Columbia University) for their support and feedback throughout this project. This research was developed as part of the interdisciplinary seminar on Anonymity & Privacy at Columbia Law School and Columbia Engineering. We would also like to thank Professor John C. Coffee, Jr. (Columbia Law School) for insightful discussions on insider trading doctrine and the development of the paper’s legal arguments.

Prompt Templates

We define our multi-stage prompting architecture in Table 7, which outlines the system instructions and the dimensional rubrics provided to the model.

TypePrompt Content
SystemYou are a securities analyst and computational linguist specializing in the analysis of SEC regulatory filings. You assess economic similarity between public companies based solely on their formal disclosures, applying the standard a sophisticated reasonable investor would use when evaluating whether two firms occupy the same market niche.
Stage 1I have uploaded the 10-K Item 7 (MD&A) for [TARGET_COMPANY], covering fiscal year [YEAR]. This was the primary disclosure available when [ACQUIRER] announced its acquisition of [TARGET_COMPANY] in [MONTH YEAR]. Using only this document, identify 10 publicly traded companies that were most similar to [TARGET_COMPANY] as of [MONTH YEAR]. Constraints: - Use No Hindsight: Base your assessment only on the state of the market as of [MONTH YEAR] - Reasonable Insider Standard: Prioritize firms that an employee at [TARGET_COMPANY] would realistically view as a peer or benchmark company - Do not include [ACQUIRER] in your results - SEC Filing Requirement: All 10 companies must have an active SEC filing obligation (10-K or 20-F). - Price Data Availability: All 10 companies must have been actively traded on a major US exchange (NYSE, NASDAQ). For each, provide: A) Name/Ticker; B) Dimensional Justification (B1: Technical, B2: Strategic, B3: Risk); C) Validation Document.
Stage 2Using only the content of these uploaded documents, assign each candidate a Total Similarity Score (0.00 to 1.00) based on: A) Technical/Product Overlap (40%): Core technology or product engines. B) Financial/Strategic Stage (30%): Revenue growth and commercialization milestones. C) Risk & Market Exposure (30%): Geographic focus and competitive headwinds. Additionally, apply dual-track analysis: D) Noun-Based Technical Comparison (HP Logic): Identify specific mechanism-of-action nouns. E) Semantic Bridging: Resolve instances where companies describe identical strategies using different terminology. This is critical to assessing the “Administrability Gap.”

Table 7. LLM Prompt Specifications.

#Company (Ticker)B1: TechnicalB2: StrategicB3: Risk
1Incyte (INCY)Targeted oncology and hematologic malignancies overlap (e.g., lymphoma, myeloma).Single-product revenue model with major global collaboration (e.g., Jakafi/Novartis).Exposure to gross-to-net pricing dynamics and reliance on ex-US partners.
2Exelixis (EXEL)Small-molecule oncology therapies overlapping prostate and breast cancer.Transition to fully integrated commercial model with internal oncology sales force.Inventory volatility and R&D expense tied to late-stage trials.
3Tesaro (TSRO)PARP inhibitor targeting ovarian and breast cancer; direct overlap with MDV3800.Scaling SG&A for commercial launch; growth via in-licensing.Manufacturing scale-up risks for PARP inhibitors.
4Clovis Oncology (CLVS)PARP inhibitor (rucaparib) for ovarian and prostate cancers.Focus on genetically defined populations using companion diagnostics.FDA setbacks and clinical/regulatory volatility.
5Seattle Genetics (SGEN)Antibody-based oncology therapies for lymphoma and breast cancer.Single-product revenue base reinvested into label expansion; co-development partnerships.Pricing pressure and complex collaboration revenue accounting.
6Ariad Pharmaceuticals (ARIA)Targeted therapies for hematologic and solid tumors.Transition to independent commercialization with specialized sales teams.High exposure to drug pricing pressure and advocacy scrutiny.
7Puma Biotechnology (PBYI)Targeted breast cancer therapy (neratinib) overlapping TNBC strategy.Growth via in-licensed assets from large pharma.Large contingent liabilities tied to milestone payments.
8Merrimack (MACK)Targeted oncology therapies for solid tumors (breast, pancreatic).Early-stage commercialization with ex-US licensing partnerships.Manufacturing scale-up and clinical trial cost expansion risks.
9Immunomedics (IMMU)Antibody therapies targeting TNBC; direct indication overlap.Advancing assets into Phase 3 with companion diagnostics.Regulatory volatility and clinical trial design risk.
10Celldex Therapeutics (CLDX)Antibody-based immuno-oncology therapies for breast cancer and lymphoma.Mixed internal discovery and in-licensing strategy.High IPR&D impairment risk and clinical failure exposure.

Table 8. Stage 1 Output: Pfizer/Medivation

RankCompanyTechStageRiskTotalSemantic Justification
1TSRO0.380.250.250.88Identical PARP inhibitor terminology and breast/ovarian cancer focus; overlapping PD-1 immuno-oncology pipeline narrative.
2EXEL0.300.280.280.86Matching U.S. profit-sharing and ex-U.S. royalty structures; identical mCRPC clinical focus.
3INCY0.250.290.280.82Single-product revenue with ex-U.S. partner; identical Medicare Part D gross-to-net pricing dynamics.
4CLVS0.350.200.250.80Direct PARP inhibitor overlap; mirrored IPR&D impairment and FDA-related clinical setbacks.
5SGEN0.250.280.250.78Identical U.S./ex-U.S. commercialization split and healthcare pricing pressure exposure.
6ARIA0.200.250.250.70Similar reliance on third-party data and exposure to global pricing and reimbursement pressures.
7MACK0.200.220.200.62Parallel ex-U.S. licensing strategy used to fund internal oncology R&D.
8CLDX0.300.100.150.55Immuno-oncology and TNBC overlap; weaker commercial-stage alignment.
9PBYI0.250.150.150.55Breast cancer focus with similar build-out of internal oncology sales infrastructure.
10IMMU0.280.100.150.53Strong TNBC overlap but penalized for pre-commercial financial stage.

Table 9. Stage 2 Output: Pfizer/Medivation

#Company (Ticker)B1: TechnicalB2: StrategicB3: Risk
1Genmab (GMAB)Direct overlap in ADCs and monoclonal antibodies; co-development of TIVDAK targeting tissue factor.50/50 co-commercialization and profit-sharing structure identical to Seagen.Identical product-specific risks (TIVDAK adoption, eye toxicity management, global rollout).
2Exelixis (EXEL)Overlap in targeted oncology; ADC pipeline targeting tissue factor (XB002).Large-scale commercial oncology firm with $1B+ revenue and global partnerships. | Exposure to U.S. pricing pressure (Medicaid rebates, 340B) and oncology competition.
3ADC Therapeutics (ADCT)Pure-play ADC company targeting hematologic malignancies (e.g., CD30-expressing lymphomas).Recently transitioned to commercial-stage biotech with FDA accelerated approvals.Manufacturing complexity and geographic commercialization via partners.
4ImmunoGen (IMGN)Core ADC platform using monoclonal antibodies to deliver cytotoxic payloads.Hybrid revenue model (product + licensing/royalties) similar to Seagen.ADC-specific manufacturing, safety, and clinical execution risks.
5MacroGenics (MGNX)Monoclonal antibodies, bispecifics, and ADCs targeting HER2-positive cancers.Early-commercial oncology firm with licensing and internal pipeline development.Competitive pressure in HER2 oncology and commercialization cost burden.
6Blueprint Medicines (BPMC)Targeted oncology therapies (kinase inhibitors); overlap via precision oncology.Commercial-stage biotech with expanding global infrastructure.Medicare/340B pricing exposure and specialty distribution dependence.
7Zymeworks (ZYME)HER2-targeted ADCs and bispecific antibodies overlapping Seagen’s pipeline.Pre-commercial with heavy reliance on licensing deals and milestone funding.Competitive dynamics in HER2 space and pipeline execution risk.
8Incyte (INCY)Broad oncology pipeline (JAK inhibitors, antibodies); limited ADC overlap.Large commercial biotech with significant royalty revenue streams.Pricing reform exposure (IRA), global partner reliance, and reimbursement risk.
9Mersana Therapeutics (MRSN)ADC platform using proprietary payload technologies (e.g., microtubule inhibitors).Pre-commercial biotech relying on partnerships to fund R&D.High clinical risk and early-stage financial uncertainty.
10Sutro Biopharma (STRO)Next-gen ADCs with linker-warhead architecture; strong platform similarity.Early-stage firm focused on partnerships and pipeline development.Manufacturing scale-up and pre-commercial execution risk.

Table 10. Stage 1 Output: Pfizer / Seagen (2023)

RankCompanyTechStageRiskTotalSemantic Justification
1GMAB0.380.260.280.92Exact match via TIVDAK co-development; identical product, revenue structure, and geographic commercialization risk profile.
2EXEL0.300.280.280.86Shared targeted oncology focus; identical tissue-factor ADC targeting and nearly identical gross-to-net pricing exposure.
3ADCT0.380.180.260.82Pure-play ADC model targeting hematologic malignancies; strong payload-delivery semantic alignment with Seagen.
4IMGN0.380.150.250.78Identical “antibody-drug conjugate” architecture; payload terminology bridges directly to Seagen’s cell-killing mechanism.
5MGNX0.340.150.220.71HER2 oncology overlap with similar commercialization dynamics and competitive pressures.
6BPMC0.250.200.250.70Shared oral oncology commercialization risks (Medicare Part D, 340B) despite different molecular modality.
7ZYME0.350.120.200.67Direct HER2-targeted ADC overlap but weaker commercial-stage alignment.
8INCY0.150.250.260.66Strong financial structure similarity (royalties, partnerships) despite technical divergence.
9MRSN0.360.080.150.59ADC payload equivalence but large gap due to pre-commercial financial profile.
10STRO0.350.080.150.58Platform-level ADC similarity (“linker-warhead”) but minimal commercialization overlap.

Table 11. Stage 2 Output: Pfizer / Seagen (2023)

Per-Event Rank Correlations

Table 12 reports the Spearman rank correlation between the LLM similarity score and the announcement-day abnormal return for each of the 30 events, computed over the per-event data in Appendices C and D. These statistics are a deterministic function of the similarity scores and abnormal returns already reported; no additional market data was collected to produce them.

IndustryEventPeersρ\rhoPerm. ppLabelSign agrees
BiopharmaceuticalsAbbVie / ImmunoGen8+0.8100.022Supports✓
BMS / Karuna8−0.2380.582Supports×
BMS / Mirati7+0.1790.713Contradicts×
BMS / MyoKardia8+0.4520.267Contradicts×
BMS / Turning Point9+0.2330.551Supports✓
Gilead / Forty Seven7−0.1430.783Contradicts✓
Merck / Acceleron7+0.1980.667Contradicts×
Merck / Pandion7+0.2140.662Supports✓
Pfizer / Medivation5−0.1030.900Supports×
Pfizer / Seagen7−0.4290.354Contradicts✓
Consumer StaplesAmazon / Whole Foods9+0.4170.269Supports✓
Danone / WhiteWave6+0.0291.000Supports✓
J.M. Smucker / Hostess5+0.1000.950Supports✓
Kroger / Albertsons9+0.2330.553Contradicts×
Reckitt / Mead Johnson9+0.0001.000Contradicts—
TechnologyAMD / Xilinx9−0.1170.776Contradicts✓
Avago / Broadcom5−0.8000.133Contradicts✓
Cisco / Splunk10+0.1880.605Mixed—
Microchip / Microsemi6−0.1430.803Mixed—
SS&C / DST Systems7−0.3210.498Contradicts✓
AutomotivesApollo / Tenneco9−0.0750.852Mixed—
Goodyear / Cooper9+0.1170.775Supports✓
Tenneco / Federal-Mogul7+0.2860.556Supports✓
Warner / Delphi7−0.5000.267Mixed—
Warner / Remy8+0.5990.125Supports✓
FinanceBB&T / SunTrust7−0.2500.595Contradicts✓
KeyCorp / First Niagara5+0.5000.450Supports✓
People’s / United6−0.3710.497Contradicts✓
Sterling / Astoria5+0.1000.950Supports✓
WSFS / Beneficial6+0.2000.714Supports✓
Mean+0.04619/25

Table 12. Per-event Spearman rank correlation between LLM similarity score and announcement-day abnormal return. ρ>0\rho> 0 indicates that more similar peers experienced higher abnormal returns (the direction the shadow trading hypothesis predicts). Permutation pp-values are exact for n≤8n \le8 and Monte Carlo (200,000 draws) otherwise. The final column reports whether the sign of ρ\rho agrees with the holistic label assigned in Table 4.

Case Studies

Reference Case: Pfizer / Medivation (2016)

The Pfizer/Medivation acquisition serves as the factual basis for [30]. Applying our pipeline ex ante, the model ranks Tesaro, Exelixis, and Incyte as the closest peers, reflecting similarities in oncology pipelines, commercialization approach, and pricing exposure.

All three top-ranked peers exhibit positive abnormal returns (+1.60%, +1.33%, +5.04%). Incyte, the firm actually traded, is ranked third and shows the strongest return among those three; the largest abnormal return in the event belongs to Puma Biotechnology (PBYI, +6.44%), ranked ninth. Consequently the rank correlation for this event is slightly negative (ρ=−0.10\rho= -0.10). This case shows that textual similarity can surface a firm the SEC regarded as economically linked, but it does not validate the pipeline: see Section 4.1 for the caveats, including the possibility that the model recalls this widely reported case from pretraining.

Tables 8 and 9 show the two-stage LLM outputs. Tables 13 and 14 summarize the abnormal return calculations and the combined similarity-ranking analysis, respectively, while Figure 3 visualizes the relationship between similarity rank and abnormal returns.

Pfizer / Medivation: Similarity Rank vs. Abnormal Return

Figure 3. Pfizer / Medivation Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
CLDX-0.012658227848101266-0.03933565933364627Decreased
EXEL0.0399999618530273440.013322530367482339Increased
INCY0.077125248975279390.050447817489734384Increased
PBYI0.091036162096226780.06435873061068177Increased
TSRO0.042723325378788240.01604589389324324Increased

Table 13. Abnormal Returns Medivation

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1TSRO0.880.04272332537878820.016045893932432Increased0.014120386626054016
2EXEL0.860.03999996185302730.0133225303674823Increased0.011457376116034779
3INCY0.820.07712524897527930.0504478174897343Increased0.041367210341582125
8CLDX0.55-0.0126582278481012-0.0393356593336462Decreased-0.021634612633505412
9PBYI0.550.09103616209622670.0643587306106817Increased0.03539730183587494

Table 14. Final Analysis Medivation

Contradicting Case: Pfizer / Seagen (2023)

The Pfizer/Seagen acquisition provides a direct counterexample in the same industry. Using the identical pipeline, the model again identifies highly similar oncology peers with comparable technical and financial profiles.

However, the market response diverges: the most similar firms show negligible or negative abnormal returns (≈0%\approx0\%, -, -), and the overall relationship between similarity and returns reverses. Despite strong textual similarity, no clear trading signal emerges, contradicting the predictability assumed by the shadow trading theory.

Tables 10 and 11 show the two-stage LLM outputs. Tables 15 and 16 summarize the abnormal return calculations and the combined similarity-ranking analysis, respectively, while Figure 4 visualizes the relationship between similarity rank and abnormal returns.

Pfizer/Seagen: Similarity Rank vs. Abnormal Return

Figure 4. Pfizer / Seagen Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
ADCT-0.07299271787868707-0.10207023020129924Decreased
EXEL0.0077937145316635875-0.02128379779094858Decreased
GMAB0.032970116856816620.003892604534204449Increased
INCY0.01624771745919238-0.012829794863419789Decreased
MGNX0.105263143742467080.0761856314198549Increased
STRO0.058479533033456980.02940202071084481Increased
ZYME0.0403531120540287160.011275599731416546Increased

Table 15. Abnormal Returns Seagen

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1GMAB0.920.03297011685681660.0038926045342044Increased0.003581196171468048
2EXEL0.860.0077937145316635-0.0212837977909485Decreased-0.01830406610021571
3ADCT0.82-0.0729927178786870-0.1020702302012992Decreased-0.08369758876506533
5MGNX0.710.10526314374246700.0761856314198549Increased0.05409179830809698
7ZYME0.670.04035311205402870.0112755997314165Increased0.007554651820049055
8INCY0.660.0162477174591923-0.0128297948634197Decreased-0.008467664609857003
10STRO0.580.05847953303345690.0294020207108448Increased0.017053172012289983

Table 16. Final Analysis Seagen

Remaining Case Studies

These subsections present the empirical case studies and results of the paper’s analysis. These sections apply the NLP pipeline to real-world acquisition events and examine whether semantic similarity between firms (based on SEC filings) actually predicts stock price reactions—i.e., whether “shadow trading” is empirically supported.

Biopharmaceuticals

ABBVIE / IMMUNOGEN

AbbVie’s acquisition of ImmunoGen is one of the clearest supporting cases in the dataset. The pipeline identifies ADC-focused peers (ADCT, STRO, MGNX) that show uniformly positive abnormal returns on announcement day, with the top-ranked firm (ADCT, +5.5% AR) exhibiting the strongest reaction. The negative trendline and positive top-3 AR direction confirm that textual similarity in this case successfully predicts the direction and magnitude of peer stock reactions, consistent with the shadow trading hypothesis.

AbbVie/ImmunoGen similarity rank versus abnormal return

Figure 5. AbbVie / ImmunoGen Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
ADCT0.077922082948612430.05530851580538361Increased
BCYC0.065088693119183470.04247512597595465Increased
CTMX0.0451127374763250840.022499170333096268Increased
GMAB0.010549869452110954-0.01206369769111786Decreased
KURA0.018967366650902547-0.0036462004923262688Decreased
MGNX0.091755326990765020.0691417598475362Increased
SNDX0.02116572005222625-0.0014478470910025652Decreased
STRO0.129870127725475240.10725656058224642Increased
ZYME0.005668956061779067-0.016944611081449748Decreased

Table 17. Abnormal Returns ImmunoGen

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1ADCT0.930.07792208294861240.0553085158053836Increased0.05143691969900675
2STRO0.870.12987012772547520.1072565605822464Increased0.09331320770655437
4MGNX0.740.09175532699076500.0691417598475362Increased0.051164902287176794
5CTMX0.680.04511273747632500.0224991703330962Increased0.015299435826505416
7BCYC0.610.06508869311918340.0424751259759546Increased0.025909826845332305
8ZYME0.520.0056689560617790-0.0169446110814497Decreased-0.008811197762353845
9GMAB0.400.0105498694521109-0.0120636976911178Decreased-0.00482547907644712
10KURA0.350.0189673666509025-0.0036462004923262Decreased-0.00127617017231417

Table 18. Final Analysis ImmunoGen

BMS / Karuna Therapeutics

The BMS/Karuna Therapeutics case is classified as supporting, driven primarily by the top-ranked firm RVPH (+17.0% AR), an extreme positive outlier. The remaining peers show mixed or near-zero reactions. This case illustrates the outlier dominance failure mode: the negative trendline is real but largely attributable to one firm rather than a systematic pattern across the cohort.

BMS / Karuna Therapeutics: Similarity Rank vs. Abnormal Return

Figure 6. BMS / Karuna Therapeutics Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
ACAD0.06368560743948670.02884434523941582Increased
AVXL0.02285712105887277-0.011984141141198112Decreased
AXSM0.059809587247459210.02496832504738833Increased
BTAI0.02020200074126503-0.014639261458805852Decreased
NBIX0.060168315168288720.025327052968217842Increased
PRTA0.0536624624479547650.018821200247883883Increased
RLMD0.0372880994297593440.002446837229688463Increased
RVPH0.205194844828023360.17035358262795247Increased

Table 19. Abnormal Returns Karuna Therapeutics

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1RVPH0.940.20519484482802330.1703535826279524Increased0.16013236767027522
2BTAI0.870.0202020007412650-0.0146392614588058Decreased-0.012736157469161045
4RLMD0.840.03728809942975930.0024468372296884Increased0.002055343272938256
5AVXL0.810.0228571210588727-0.0119841411411981Decreased-0.009707154324370462
6PRTA0.760.05366246244795470.0188212002478838Increased0.014304112188391688
7ACAD0.730.06368560743948670.0288443452394158Increased0.021056372024773534
8AXSM0.720.05980958724745920.0249683250473883Increased0.017977194034119577
10NBIX0.630.06016831516828870.0253270529682178Increased0.015956043369977215

Table 20. Final Analysis Karuna Therapeutics

BMS / MYOKARDIA

Despite a negative trendline, BMS/MyoKardia is classified as strongly contradicting. The top-ranked firm CYTK (+14.3% AR) is a large positive outlier, while all remaining high-similarity peers show negative or negligible ARs. This is a textbook outlier dominance case: the trendline direction is mechanically driven by a single observation and does not reflect a consistent relationship between similarity and peer stock reactions.

Bristol Myers Squibb/MyoKardia: Similarity Rank vs. Abnormal Return

Figure 7. BMS / MyoKardia Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
BBIO0.004292979550332126-0.04509094642333769Decreased
CYCN0.04213942069154834-0.00724450528212147Decreased
CYTK0.192580278587391520.1431963526137217Increased
FOLD0.066072707647002310.016688781673332494Increased
LXRX-0.02068971697413619-0.070073642947806Decreased
MIST0.02215658726588405-0.027227338707785764Decreased
RCKT0.0753056913874760.02592176541380619Increased
TENX-0.08759124087591241-0.13697516684958222Decreased

Table 21. Abnormal Returns Myokardia

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1CYTK0.940.19258027858739150.1431963526137217Increased0.13460457145689841
2CYCN0.800.0421394206915483-0.0072445052821214Decreased-0.00579560422569712
4MIST0.690.0221565872658840-0.0272273387077857Decreased-0.01878686370837213
5RCKT0.630.07530569138747600.0259217654138061Increased0.016330712210697842
6LXRX0.58-0.0206897169741361-0.0700736429478060Decreased-0.04064271290972748
7TENX0.50-0.0875912408759124-0.1369751668495822Decreased-0.0684875834247911
8BBIO0.420.0042929795503321-0.0450909464233376Decreased-0.018938197497801793
10FOLD0.300.06607270764700230.0166887816733324Increased0.00500663450199972

Table 22. Final Analysis Myokardia

BMS / TURNING POINT THERAPEUTICS

BMS/Turning Point Therapeutics is a clean supporting case. The top three peers (RLAY, RVMD, NUVL) all show positive abnormal returns, and the negative trendline is consistent across the cohort. This case, alongside AbbVie/ImmunoGen and Merck/Pandion, is among the clearest instances in which 10-K similarity does track peer reactions in a targeted oncology context—though only AbbVie/ImmunoGen reaches individual significance (Table 12).

BMS / Turning Point Therapeutics: Similarity Rank vs. Abnormal Return

Figure 8. BMS / Turning Point Therapeutics Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
BDTX0.02873560462197694-0.006745240120843793Decreased
COGT-0.018140572896618242-0.05362141763943898Decreased
DAWN0.0524737482202607340.01699290347744Increased
ERAS0.132264600104655740.096783755361835Increased
KURA0.005930313199658817-0.029550531543161916Decreased
NUVL0.105896403890577880.07041555914775714Increased
RLAY0.04469606491714230.009215220174321567Increased
RVMD0.105780347075725710.07029950233290498Increased
TYRA0.076134697911439260.040653853168618526Increased

Table 23. Abnormal Returns Turning Point Therapeutics

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1RLAY0.920.04469606491714230.0092152201743215Increased0.00847800256037578
2RVMD0.910.10578034707572570.0702995023329049Increased0.06397254712294347
3NUVL0.900.10589640389057780.0704155591477571Increased0.0633740032329814
4BDTX0.870.0287356046219769-0.0067452401208437Decreased-0.005868358905134019
5TYRA0.830.07613469791143920.0406538531686185Increased0.03374269812995335
6COGT0.82-0.0181405728966182-0.0536214176394389Decreased-0.0439695624643399
7ERAS0.810.13226460010465570.0967837553618350Increased0.07839484184308636
8DAWN0.770.05247374822026070.0169929034774400Increased0.013084535677628802
9KURA0.760.0059303131996588-0.0295505315431619Decreased-0.022458403972803045

Table 24. Final Analysis Turning Point Therapeutics

BMS / MIRATI THERAPEUTICS

BMS/Mirati Therapeutics is classified as contradicting despite a negative trendline. The top-3 AR direction is mixed (++, −-, ++), and the second-ranked firm (TNGX, −12.5%-12.5\% AR) shows a large negative reaction that disrupts any clean positive signal. The negative trendline is driven by the interaction between RVMD (high similarity, positive AR) and TNGX (high similarity, strongly negative AR), producing a noisy, uninterpretable pattern overall.

Bristol Myers Squibb / Mirati Therapeutics: Similarity Rank vs. Abnormal Return

Figure 9. BMS / Mirati Therapeutics Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
ERAS-0.0045248824844553250.0033664856347955076Increased
EXEL-0.03285326574234348-0.024961897623092644Decreased
IDYA0.002999622177622860.010890990296873692Increased
PRLD-0.09665426949169337-0.08876290137244254Decreased
RLAY0.0301568155790580480.03804818369830888Increased
RVMD0.0171724593492527460.02506382746850358Increased
TNGX-0.13268894672203788-0.12479757860278705Decreased

Table 25. Abnormal Returns Mirati Therapeutics

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1RVMD0.910.01717245934925270.0250638274685035Increased0.022808082996338186
2TNGX0.85-0.1326889467220378-0.1247975786027870Decreased-0.10607794181236894
3IDYA0.840.00299962217762280.0108909902968736Increased0.009148431849373823
5ERAS0.80-0.00452488248445530.0033664856347955Increased0.0026931885078364
6RLAY0.780.03015681557905800.0380481836983088Increased0.029677583284680863
9PRLD0.58-0.0966542694916933-0.0887629013724425Decreased-0.051482482796016645
10EXEL0.45-0.0328532657423434-0.0249618976230926Decreased-0.01123285393039167

Table 26. Final Analysis Mirati Therapeutics

Gilead / Forty Seven

Gilead/Forty Seven is classified as contradicting. The trendline is positive (higher similarity correlates with lower ARs), and the top-ranked firm shows a negative abnormal return. The case is notable because IPHA, ranked ninth, shows the largest positive AR (+9.2%), inverting the expected pattern entirely.

Gilead / Forty Seven: Similarity Rank vs. Abnormal Return

Figure 10. Gilead / Forty Seven Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
ALEC0.03384280642895883-0.005628199429125465Decreased
CLDX-0.015209200962377005-0.0546802068204613Decreased
FATE0.061986281460637210.02251527560255292Increased
IPHA0.13099036179599110.0915193559379068Increased
KURA0.0405965813737839340.0011255755156996394Increased
SNDX0.0414453081452503740.00197430228716608Increased
TGTX0.115537831651636990.0760668257935527Increased

Table 27. Abnormal Returns Forty Seven

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
3CLDX0.84-0.0152092009623770-0.0546802068204613Decreased-0.045931373729187486
4TGTX0.800.11553783165163690.0760668257935527Increased0.06085346063484216
5SNDX0.750.04144530814525030.0019743022871660Increased0.0014807267153745
7FATE0.690.06198628146063720.0225152756025529Increased0.0155355401657615
8KURA0.590.04059658137378390.0011255755156996Increased0.0006640895542627639
9IPHA0.580.13099036179599110.0915193559379068Increased0.05308122644398594
10ALEC0.550.0338428064289588-0.0056281994291254Decreased-0.00309550968601897

Table 28. Final Analysis Forty Seven

Merck / Acceleron Pharma

Merck/Acceleron Pharma is strongly contradicting. Despite a negative trendline, the top-ranked peers show near-zero or negative ARs, while the two firms with the highest and only positive ARs (SRRK, PTGX) rank in the middle of the similarity distribution. The announcement-day signal is weak across the cohort and does not concentrate in the most similar firms.

Merck / Acceleron Pharma Similarity Rank vs. Abnormal Return

Figure 11. Merck / Acceleron Pharma Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
AGIO-0.0012983674482242133-0.0019350482482927741Decreased
GOSS-0.04410645935599469-0.04474314015606325Decreased
KROS-0.03770370260394145-0.03834038340401001Decreased
LQDA-0.010752678064499305-0.011389358864567866Decreased
PTGX0.0085372578904745460.007900577090405985Increased
SRRK0.053269539265493740.05263285846542518Increased
UTHR-0.004959558862238881-0.005596239662307442Decreased

Table 29. Abnormal Returns Acceleron Pharma

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
2AGIO0.70-0.0012983674482242-0.0019350482482927Decreased-0.00135453377380489
3KROS0.68-0.0377037026039414-0.0383403834040100Decreased-0.0260714607147268
4SRRK0.650.05326953926549370.0526328584654251Increased0.03421135800252632
5PTGX0.600.00853725789047450.0079005770904059Increased0.00474034625424354
6GOSS0.60-0.0441064593559946-0.0447431401560632Decreased-0.026845884093637917
7UTHR0.50-0.0049595588622388-0.0055962396623074Decreased-0.0027981198311537
9LQDA0.40-0.0107526780644993-0.0113893588645678Decreased-0.00455574354582712

Table 30. Final Analysis Acceleron Pharma

Merck / Pandion Therapeutics

Merck/Pandion Therapeutics is a strongly supporting case. All top-ranked peers show positive abnormal returns, and the pattern holds across the full cohort: six of seven firms show positive ARs, with the negative case (PTGX) ranking sixth. This is one of the most consistent results in the dataset and suggests that in immuno-oncology acquisitions with tightly defined peer groups, textual similarity can reliably predict announcement-day reactions.

Merck / Pandion Therapeutics: Similarity Rank vs. Abnormal Return

Figure 12. Merck / Pandion Therapeutics Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
ANAB-0.018882175553005060.021744613828727936Increased
CUE-0.0104712299567009030.03015555942503209Increased
EQ-0.034351142028007810.006275647353725183Increased
KZR-0.0228873108190524520.017739478562680544Increased
NKTR-0.0144927005310085590.026134088850724435Increased
PTGX-0.052153288134218864-0.011526498752485868Decreased
XNCR-0.035876323267356640.004750466114376357Increased

Table 31. Abnormal Returns Pandion Therapeutics

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1CUE0.94-0.01047122995670090.0301555594250320Increased0.028346225859530078
3ANAB0.83-0.01888217555300500.0217446138287279Increased0.018048029477844157
5XNCR0.80-0.03587632326735660.0047504661143763Increased0.0038003728915010403
6PTGX0.79-0.0521532881342188-0.0115264987524858Decreased-0.009105934014463782
7EQ0.78-0.03435114202800780.0062756473537251Increased0.004895004935905578
8NKTR0.76-0.01449270053100850.0261340888507244Increased0.019861907526550544
10KZR0.72-0.02288731081905240.0177394785626805Increased0.01277242456512996

Table 32. Final Analysis Pandion Therapeutics

Consumer Staples

AMAZON / WHOLE FOODS

Amazon/Whole Foods is classified as supporting despite all peers showing negative ARs. The key is the trendline: higher-similarity grocery peers (SFM, NGVC) decline more than lower-similarity peers, which is consistent with the shadow trading mechanism in reverse—the acquisition of a major competitor signals competitive pressure rather than acquisition premium for peers. The negative trendline remains directionally consistent with the hypothesis.

Amazon/Whole Foods similarity rank versus abnormal return

Figure 13. Amazon / Whole Foods Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
COST-0.07192040374481146-0.06796522838659612Decreased
IMKTA-0.03000031493783293-0.026045139579617585Decreased
KR-0.09242663874926776-0.08847146339105243Decreased
NGVC-0.026755852092367225-0.022800676734151884Decreased
SFM-0.06289026951890986-0.058935094160694514Decreased
TGT-0.05138819614229597-0.047433020784080626Decreased
UNFI-0.10968554994595126-0.10573037458773592Decreased
VLGEA-0.000399507699949629140.003555667658265714Increased
WMK-0.04752578168203486-0.04357060632381952Decreased

Table 33. Abnormal Returns Whole Foods

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1SFM0.92-0.0628902695189098-0.0589350941606945Decreased-0.05422028662783894
2NGVC0.90-0.0267558520923672-0.0228006767341518Decreased-0.020520609060736623
3VLGEA0.83-0.00039950769994960.0035556676582657Increased0.0029512041563605307
4KR0.76-0.0924266387492677-0.0884714633910524Decreased-0.06723831217719982
5IMKTA0.65-0.0300003149378329-0.0260451395796175Decreased-0.016929340726751375
6WMK0.60-0.0475257816820348-0.0435706063238195Decreased-0.026142363794291697
8UNFI0.50-0.1096855499459512-0.1057303745877359Decreased-0.05286518729386795
9TGT0.45-0.0513881961422959-0.0474330207840806Decreased-0.02134485935283627
10COST0.35-0.0719204037448114-0.0679652283865961Decreased-0.023787829935308636

Table 34. Final Analysis Whole Foods

Danone / WhiteWave Foods

Danone/WhiteWave Foods is a supporting case. The top-ranked firm (HAIN, +6.9% AR) shows a clear positive reaction, and the negative trendline holds moderately across the cohort. Several mid- and lower-ranked firms show slightly negative but negligible reactions, consistent with a weaker similarity-return relationship in consumer staples compared to biopharmaceuticals.

Danone/WhiteWave similarity rank versus abnormal return

Figure 14. Danone / WhiteWave Foods Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
BGS-0.001464230372955497-0.0031091979285966117Decreased
CVGW-0.004072587271339847-0.005717554826980962Decreased
DFP0.00041192300467115514-0.0012330445509699595Decreased
GIS-0.0008324021591602719-0.0024773697148013866Decreased
HAIN0.070854534164777980.06920956660913687Increased
LWAY0.087523329058669980.08587836150302887Increased
POST0.00354735286684035830.0019023853111992436Increased

Table 35. Abnormal Returns WhiteWave Foods

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1HAIN0.870.07085453416477790.0692095666091368Increased0.060212322949949014
3GIS0.76-0.0008324021591602-0.0024773697148013Decreased-0.0018828009832489879
6POST0.690.00354735286684030.0019023853111992Increased0.001312645864727448
7CVGW0.68-0.0040725872713398-0.0057175548269809Decreased-0.0038879372823470123
8BGS0.64-0.0014642303729554-0.0031091979285966Decreased-0.001989886674301824
9LWAY0.630.08752332905866990.0858783615030288Increased0.054103367746908146

Table 36. Final Analysis WhiteWave Foods

JM SMUCKER / HOSTESS

JM Smucker/Hostess is a clean supporting case. All five peers for which data was available show positive abnormal returns, and the top three (FLO, JJSF, THS) show the largest gains, consistent with the shadow trading hypothesis. This is one of the strongest results in the consumer staples subsample.

JM Smucker / Hostess: Similarity Rank vs. Abnormal Return

Figure 15. JM Smucker / Hostess Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
CPB0.0128025468888891690.00263678770741119Increased
FLO0.0143105565285681340.004144797347090156Increased
GIS-0.003930505467713301-0.01409626464919128Decreased
HSY0.014351924800572750.004186165619094772Increased
JJSF0.0213976063910744050.011231847209596427Increased
POST0.004609905973885611-0.005555853207592367Decreased
THS0.0336856595260771750.023519900344599196Increased

Table 37. Abnormal Returns Hostess

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1FLO0.920.01431055652856810.0041447973470901Increased0.0038132135593228924
2JJSF0.860.02139760639107440.0112318472095964Increased0.009659388600252904
3THS0.820.03368565952607710.0235199003445991Increased0.019286318282571262
5CPB0.780.01280254688888910.0026367877074111Increased0.002056694411780658
7HSY0.720.01435192480057270.0041861656190947Increased0.003014039245748184

Table 38. Final Analysis Hostess

Kroger / Albertsons

Kroger/Albertsons is classified as contradicting. Despite a negative trendline, the top-3 AR direction is mixed (+, −, +), and the overall distribution shows no consistent relationship between similarity rank and return magnitude. Regulatory uncertainty surrounding the proposed merger likely introduced idiosyncratic noise that overwhelmed any fundamental peer signal.

Kroger/Albertsons: Similarity Rank vs. Abnormal Return

Figure 16. Kroger / Albertsons Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
BJ-0.0163412314102842450.002392864995054235Increased
COST-0.02850488629098535-0.00977078988564687Decreased
GO-0.07023213268824618-0.05149803628290771Decreased
IMKTA-0.055493940265775386-0.0367598438604369Decreased
SFM-0.06223925247850891-0.04350515607317043Decreased
TGT-0.05469759121142679-0.035963494806088306Decreased
VLGEA-0.016496765407659680.0022373309976788Increased
WMK-0.04363704397423403-0.024902947568895547Decreased
WMT-0.0139855917184197740.0047485046869187066Increased

Table 39. Abnormal Returns Albertsons

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1WMK0.94-0.0436370439742430-0.0249029475688955Decreased-0.023408770714761766
2IMKTA0.89-0.0554939402657753-0.0367598438604369Decreased-0.03271626103578884
3VLGEA0.88-0.01649676540765960.0022373309976788Increased0.001968851277957344
4WMT0.76-0.01398559171841970.0047485046869187Increased0.003608863562058212
5SFM0.68-0.0622392524785089-0.0435051560731704Decreased-0.029583506129755875
6COST0.65-0.0285048862909853-0.0097707898856468Decreased-0.00635101342567042
7BJ0.63-0.01634123141028420.0023928649950542Increased0.001507504946884146
8TGT0.58-0.0546975912114267-0.0359634948060883Decreased-0.020858826987531214
9GO0.45-0.0702321326882461-0.0514980362829077Decreased-0.023174116327308467

Table 40. Final Analysis Albertsons

Reckitt Benckiser / Mead Johnson Nutrition

Reckitt Benckiser/Mead Johnson is contradicting. The trendline is positive (primarily due to the outlier), but the top-ranked firm (ABT) shows essentially zero abnormal return. The acquisition of an infant nutrition company does not appear to have generated meaningful peer firm reactions, suggesting limited economic linkage within the consumer staples peer group as identified from 10-K text.

Reckitt Benckiser/Mead Johnson Nutrition: Similarity Rank vs. Abnormal Return

Figure 17. Reckitt Benckiser / Mead Johnson Nutrition Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
ABT0.00470164327620931-6.045667372484714e-05Decreased
CPB-0.008928184876190674-0.013690284826124832Decreased
GIS-0.004757503745379787-0.009519603695313943Decreased
HAIN-0.0028468067310845843-0.007608906681018742Decreased
HLF0.0160907985807568940.011328698630822736Increased
KHC0.004677904408272617-8.419554166153986e-05Decreased
KMB-0.00024270730659042898-0.005004807256524586Decreased
MDLZ0.0064878019838111180.001725702033876961Increased
PRGO-0.0005037224269522213-0.005265822376886378Decreased

Table 41. Abnormal Returns Mead Johnson Nutrition

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1ABT0.880.0047016432762093-6.045667372484714e-05Decreased-5.320187287786549e-05
2PRGO0.80-0.0005037224269522-0.0052658223768863Decreased-0.00421265790150904
3HAIN0.60-0.0028468067310845-0.0076089066810187Decreased-0.00456534400861122
4KMB0.55-0.0002427073065904-0.0050048072565245Decreased-0.0027526439910884755
5KHC0.490.0046779044082726-8.419554166153986e-05Decreased-4.125581541415453e-05
6MDLZ0.420.00648780198381110.0017257020338769Increased0.000724794854228298
7GIS0.38-0.0047575037453797-0.0095196036953139Decreased-0.003617449404219282
9CPB0.30-0.0089281848761906-0.0136902848261248Decreased-0.00410708544783744
10HLF0.270.01609079858075680.0113286986308227Increased0.003058748630322129

Table 42. Final Analysis Mead Johnson Nutrition

Technology

AMD / XILINX

AMD/Xilinx is contradicting. Most semiconductor peers show negative ARs and the top three peers are inconsistent. The positive trendline indicates that lower-similarity firms slightly outperformed higher-similarity ones, inverting the shadow trading prediction. Semiconductor M&A events may reflect competitive consolidation rather than acquisition premium contagion.

AMD/Xilinx similarity rank versus abnormal return

Figure 18. AMD / Xilinx Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
ADI-0.006934434890752879-0.011939806696974281Decreased
AVGO-0.009475118768035064-0.014480490574256464Decreased
INTC-0.02311685008217791-0.028122221888399312Decreased
LSCC0.010901672408756630.005896300602535229Increased
MCHP-0.016221798231416912-0.021227170037638313Decreased
NVDA0.0194425587072948640.014437186901073464Increased
NXPI-0.012306568731181536-0.017311940537402936Decreased
QCOM-0.0022979824548389656-0.007303354261060367Decreased
TXN0.0080161577808645250.0030107859746431242Increased

Table 43. Abnormal Returns Xilinx

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1LSCC0.860.010901672408075660.0058963006025352Increased0.005070818518180272
2INTC0.75-0.02311685008281779-0.0281222218883993Decreased-0.021091666416299476
3MCHP0.72-0.0162217982314169-0.0212271700376383Decreased-0.015283562427099575
4NVDA0.650.01944255870729480.0144371869010734Increased0.009384171148569771
5AVGO0.60-0.0094751187680350-0.0144804905742564Decreased-0.00868829434455384
6QCOM0.55-0.0022979824548389-0.0073033542610603Decreased-0.0040168448435831654
7NXPI0.50-0.0123065687311815-0.0173119405374029Decreased-0.00865597026870145
8ADI0.48-0.0069344348907528-0.0119398066969742Decreased-0.005731107214547616
9TXN0.350.00801615778086450.0030107859746431Increased0.001053775091125085

Table 44. Final Analysis Xilinx

Avago / Broadcom

Avago/Broadcom is contradicting. The trendline is positive, and the top two peers (QCOM, SWKS) both show negative ARs, while the fourth and fifth peers show extremely high ARs relative to the cohort. This case, along with AMD/Xilinx, suggests that in semiconductor acquisitions, the announcement-day signal does not propagate to textually similar peers in a manner consistent with the shadow trading hypothesis.

Avago/Broadcom similarity rank versus abnormal return

Figure 19. Avago / Broadcom Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
CAVM0.0079320849871343120.009987993662528285Increased
MXL-0.00100193142947568540.001053977245918287Increased
QCOM-0.0059139995490219-0.003858090873627927Decreased
SLAB0.0051263749445203180.007182283619914291Increased
SWKS-0.009021781906765464-0.006965873231371492Decreased

Table 45. Abnormal Returns Broadcom

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1QCOM0.91-0.0059139995490219-0.0038580908736279Decreased-0.003510862695001389
2SWKS0.83-0.0090217819067654-0.0069658732313714Decreased-0.005781674782038262
3MXL0.74-0.00100193142947560.0010539772459182Increased0.0007799431619794681
6CAVM0.600.00793208498713430.0099879936625282Increased0.005992796197516919
7SLAB0.580.00512637494452030.0071822836199142Increased0.004165724499550235

Table 46. Final Analysis Broadcom

Cisco / Splunk

Cisco/Splunk is classified as mixed. The top-3 AR direction is (−-, ++, ++). The trendline is negative, and the distribution is bimodal: some peers show meaningful positive reactions (ESTC, DT) while others decline sharply (DDOG, SNOW). The mixed classification reflects genuine ambiguity rather than a clean contradicting signal.

Cisco/Splunk similarity rank versus abnormal return scatter plot

Figure 20. Cisco / Splunk Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
CRWD-0.0082982525804894370.006947863262951159Increased
DDOG-0.03943514136301994-0.024189025519579344Decreased
DT0.0036756360853040540.01892175192874465Increased
ESTC0.018196091121278310.033442206964718904Increased
PANW-0.03340416705656666-0.018158051213126063Decreased
RPD-0.02088660657461326-0.005640490731172662Decreased
SNOW-0.060518223589588895-0.0452721077461483Decreased
S-0.038670659767480936-0.02342454392404034Decreased
TENB-0.030199183741365668-0.014953067897925072Decreased
VRNS-0.038279056262499216-0.02303294041905862Decreased

Table 47. Abnormal Returns Splunk

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1DDOG0.92-0.0394351413630199-0.0241890255195793Decreased-0.022253903478012958
2ESTC0.890.01819609112127830.0334422069647189Increased0.02976356419859982
3DT0.870.00367563608530400.0189217519287446Increased0.016461924178007802
4VRNS0.76-0.0382790562624992-0.0230329404190586Decreased-0.017505034718484535
5RPD0.74-0.0208866065746132-0.0056404907311726Decreased-0.004173963141067724
6S0.73-0.0386706597674809-0.0234245439240403Decreased-0.01709991706454942
7CRWD0.71-0.00829825258048940.0069478632629511Increased0.004932982916695281
8SNOW0.70-0.0605182235895888-0.0452721077461483Decreased-0.03169047542230381
9PANW0.63-0.0334041670565666-0.0181580512131260Decreased-0.011439572264269381
10TENB0.56-0.0301991837413656-0.0149530678979250Decreased-0.008373718022838001

Table 48. Final Analysis Splunk

Microchip / Microsemi

Microchip/Microsemi is also classified as mixed. All identified peers show small negative or positive ARs close to zero, and the trendline is positive but relatively flat when considering the scale. This case is notable for its uniformly low magnitude reactions, suggesting the acquisition was broadly anticipated by the market and generated no meaningful announcement-day peer spillover.

Microchip Technology/Microsemi: Similarity Rank vs. Abnormal Return

Figure 21. Microchip / Microsemi Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
ADI-0.020410672883296514-0.003834261127835862Decreased
AVGO-0.0127811754228502640.0037952363326103876Increased
LSCC-0.028286201301122735-0.011709789545662083Decreased
MTSI-0.0070290354947388820.00954737626072177Increased
SLAB-0.017647075143089906-0.0010706633876292546Decreased
SMTC-0.025260096590624392-0.00868368483516374Decreased

Table 49. Abnormal Returns Microsemi

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
2ADI0.87-0.0204106728832965-0.0038342611278358Decreased-0.003335807181217146
3SLAB0.84-0.0176470751430899-0.0010706633876292Decreased-0.0008993572456085279
4SMTC0.81-0.0252600965906243-0.0086836848351637Decreased-0.007033784716482597
6MTSI0.75-0.00702903549473880.0095473762607217Increased0.007160532195541275
7LSCC0.72-0.0282862013011227-0.0117097895456620Decreased-0.008431048472876639
8AVGO0.68-0.01278117542285020.0037952363326103Increased0.002580760706175004

Table 50. Final Analysis Microsemi

SS&C / DST SYSTEMS

SS&C/DST Systems is contradicting. Despite most peers showing positive ARs, the trendline is inverted: lower-similarity peers (MMS, CNDT) show the largest positive reactions, while higher-similarity peers (SEIC, FIS) show negligible returns. This is a case where broad sector sentiment lifted the cohort but failed to concentrate in the most textually similar firms.

SS&C Technologies/DST Systems similarity rank versus abnormal return

Figure 22. SS&C / DST Systems Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
BR0.0050545692987703410.0009727415307561648Increased
CNDT0.0170419217343939880.012960093966379812Increased
CTSH0.000540932537022394-0.003540895230991782Decreased
DXC-0.0013830794533293202-0.005464907221343496Decreased
EXLS0.0117206617658187130.007638833997804536Increased
FISV0.005985018298631030.0019031905306168542Increased
FIS0.00514028778227580350.0010584600142616273Increased
MMS0.027821373060979620.023739545292965444Increased
SEIC0.0077333664307573370.003651538662743161Increased

Table 51. Abnormal Returns DST Systems

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1SEIC0.870.00773336643075730.0036515386627431Increased0.003176838636586497
2FIS0.790.00514028778227580.0010584600142616Increased0.000836183411266664
3BR0.770.00505456929877030.0009727415307561Increased0.0007490109786821971
5EXLS0.640.01172066176581870.0076388339978045Increased0.00488885375859488
6CNDT0.540.01704192173439390.0129600939663798Increased0.006998450741845092
7CTSH0.450.0005409325370223-0.0035408952309917Decreased-0.0015934028539462649
8MMS0.360.02782137306097960.0237395452929654Increased0.008546236305467544

Table 52. Final Analysis DST Systems

Automobiles

APOLLO / TENNECO

Apollo/Tenneco is classified as mixed. The trendline is positive and the top-3 direction is (−-, ++, −-), reflecting the split between DAN (top-ranked, strongly negative AR) and BWA (second-ranked, strongly positive AR). The two most similar firms diverge sharply in their announcement-day reactions, making the case uninterpretable under either hypothesis.

Apollo/Tenneco: Similarity Rank vs. Abnormal Return

Figure 23. Apollo / Tenneco Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
SMP-0.05646922230037127-0.023050145168253267Decreased
GTX-0.002915413435587910.03050366369653009Increased
DCH0.069532263128273650.10295134026039165Increased
BWA-0.0016517810764135660.031767296055704435Increased
MOD-0.007952373165350150.025466703966767854Increased
DORM-0.052984267629254074-0.01956519049713607Decreased
FOXF-0.0329197957543586260.0004992813777593766Increased
CPS0.0187499818392100650.052169058971328064Increased
DAN-0.0966093471696384-0.0631902700375204Decreased
LEA-0.00316393388375055830.030255143248367443Increased

Table 53. Abnormal Returns Tenneco

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1DAN0.90-0.0966093471696384-0.0631902700375204Decreased-0.05687124303376836
2BWA0.90-0.00165178107641350.0317672960557044Increased0.02859056645013396
3SMP0.84-0.0564692223003712-0.0230501451682532Decreased-0.01936212194133269
5LEA0.74-0.00316393388375050.0302551432483674Increased0.022388806003791877
6GTX0.70-0.00291541343558790.03050366369653Increased0.021352564587571
7CPS0.690.018749981839210.052169058971328Increased0.03599665069021632
8DORM0.62-0.052984267629254-0.019565190497136Decreased-0.012130418108224319
9MOD0.55-0.00795237316535010.0254667039667678Increased0.014006687181722292
10FOXF0.50-0.03291979575435860.0004992813777593Increased0.00024964068887965

Table 54. Final Analysis Tenneco

Goodyear / Cooper

Goodyear/Cooper is a strongly supporting case. Eight of nine peers show positive abnormal returns, and the negative trendline holds clearly across the cohort. This is the most uniformly positive result in the Automotive subsample and suggests that in tire and rubber supply chains, textual similarity from 10-K disclosures reliably identifies economically linked firms.

Goodyear/Cooper: Similarity Rank vs. Abnormal Return

Figure 24. Goodyear / Cooper Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
DORM0.101525534619073890.09833821671472792Increased
TWI0.015815098566646110.012627780662300143Increased
DAN0.046229959577418030.04304264167307206Increased
CPS0.075744403594307220.07255708568996125Increased
GTES0.011183080436563930.007995762532217964Increased
MYE0.0105120608181936760.007324742913847709Increased
TEN-0.0020920732042427684-0.005279391108588736Decreased
MPAA0.0303029886014135120.027115670697067544Increased
SMP0.040166620896674920.036979302992328954Increased

Table 55. Abnormal Returns Cooper

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1TWI0.860.01581509856664610.0126277806623001Increased0.010859891369578085
2CPS0.770.07574440359430720.0725570856899612Increased0.05586895598127013
3GTES0.750.01118308043656390.0079957625322179Increased0.005996821899163425
4SMP0.600.04016662089667490.0369793029923289Increased0.02218758179539734
6DORM0.560.10152553461907380.0983382167147279Increased0.05506940136024763
7MYE0.520.01051206081819360.0073247429138477Increased0.0038088663152008044
8TEN0.50-0.0020920732042427-0.0052793911085887Decreased-0.00263969555429435
9DAN0.480.0462299595774180.043042641673072Increased0.02066046800307456
10MPAA0.400.03030298860141350.0271156706970675Increased0.010846268278827

Table 56. Final Analysis Cooper

Tenneco / Federal-Mogul

Tenneco/Federal-Mogul is classified as supporting despite a mixed top-3 direction (−-, ++, −-). The classification reflects the overall pattern: the two firms with positive reactions (BWA, rank 4; SMP, rank 8) are outweighed by larger negative reactions concentrated in the lowest-ranked peers (MOD, LKQ), producing the negative trendline. The holistic analysis supports the hypothesis even though the top-ranked firm (DAN) shows a small negative AR.

Tenneco/FDML: Similarity Rank vs. Abnormal Return

Figure 25. Tenneco / Federal-Mogul Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
BWA0.0279233689722556650.012623729497841695Increased
CMI0.034632826818968510.019333187344555454Increased
DORM0.01464721921591628-0.000652420258497691Decreased
LKQ0.0005245657554827761-0.014775073718931194Decreased
SMP0.0174573297219315260.0021576902475175554Increased
DAN0.012209253473614552-0.0030903860007994183Decreased
DCH0.0006250143051147461-0.014674625169299224Decreased
CPS0.014450176569453304-0.0008494629049606659Decreased
MOD0.004926127351136306-0.010373512123277664Decreased

Table 57. Abnormal Returns Federal

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
2DAN0.910.0122092534736145-0.0030903860007994Decreased-0.0028122512607274544
4BWA0.850.02792336897225560.0126237294978416Increased0.01073017007316536
6CPS0.820.0144501765694533-0.0008494629049606Decreased-0.0006965595820676919
7MOD0.740.0049261273511363-0.0103735121232776Decreased-0.007676398971225424
8SMP0.600.01745732972193150.0021576902475175Increased0.0012946141485104998
9DORM0.510.0146472192159162-0.0006524202584976Decreased-0.000332734331833776
10LKQ0.370.0005245657554827-0.0147750737189311Decreased-0.005466777276004507

Table 58. Final Analysis Federal

Warner / Delphi

Warner/Delphi is classified as mixed. The trendline is positive and the top-3 direction is (−-, ≈0\approx0, ++), with the top-ranked firm (TEN) showing a negative AR and the third-ranked (DAN) showing a positive one. The lack of a consistent pattern across either direction prevents a clear supporting or contradicting classification.

Warner/Delphi: Similarity Rank vs. Abnormal Return

Figure 26. Warner / Delphi Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
VC0.046621391221570090.04015559590695077Increased
DCH0.0181043738102557860.011638578495636463Increased
SMP-0.0027740168695917916-0.009239812184211115Decreased
APTV0.0117381438005276120.00527234848590829Increased
TEN-0.014577270889797867-0.02104306620441719Decreased
ST0.0106993778621449560.004233582547525633Increased
SRI0.0021413085282714493-0.004324486786347873Decreased
DAN0.0276905930959655720.021224797781346248Increased
CMI0.0079503621878417910.0014845668732224688Increased
LEA-0.03579390798828836-0.04225970330290768Decreased

Table 59. Abnormal Returns Delphi

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1TEN0.94-0.0145772708897978-0.0210430662044171Decreased-0.019780482232152074
2CMI0.870.00795036218784170.0014845668732224Increased0.001291573179703488
3DAN0.850.02769059309596550.0212247977813462Increased0.01804107811414427
4LEA0.81-0.0357939079882883-0.0422597033029076Decreased-0.034230359675355156
5SRI0.780.0021413085282714-0.0043244867863478Decreased-0.003373099693351284
7APTV0.730.01173814380052760.0052723484859082Increased0.0038488143947129857
8VC0.650.046621391221570.0401555959069507Increased0.026101137339517955

Table 60. Final Analysis Delphi

WARNER / REMY

Warner/Remy is a strongly supporting case. Seven of eight peers show positive abnormal returns, and the negative trendline holds across the cohort. The top-ranked firm (MPAA, +3.9% AR) and second-ranked firm (SMP, +1.0% AR) both show positive reactions consistent with the shadow trading hypothesis.

Warner/Remy: Similarity Rank vs. Abnormal Return

Figure 27. Warner / Remy Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
ALSN0.0113168570706656660.00024048661547815438Increased
DAN0.028527911017087530.017451540561900018Increased
LKQ0.0141774398992372260.003101069444049715Increased
MPAA0.049710556443782550.03863418598859504Increased
SRI0.0008417700997898144-0.010234600355397697Decreased
DORM0.012505258327278290.0014288878720907783Increased
CVGI0.015974425076088280.004898054620900767Increased
SMP0.0210617457274004830.009985375272212972Increased

Table 61. Abnormal Returns Remy

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1MPAA0.940.04971055644378250.0386341859885950Increased0.0363161348292793
2SMP0.870.02106174572740040.0099853752722129Increased0.008687276486825221
3DORM0.710.01250525832727820.0014288878720907Increased0.0010145103891843969
4LKQ0.700.01417743989923720.0031010694440497Increased0.00217074861083479
7DAN0.590.02852791101708750.0174515405619000Increased0.0102964089315210
8ALSN0.560.01131685707066560.0002404866154781Increased0.00013467250466773603
9CVGI0.480.01597442507608820.0048980546209007Increased0.0023510662180323363
10SRI0.480.0008417700997898-0.0102346003553976Decreased-0.004912608170590848

Table 62. Final Analysis Remy

Finance

BB&T / SunTrust

BB&T/SunTrust presents a nuanced case. All peer firms show positive abnormal returns on announcement day, which is consistent with the weak form of the shadow trading hypothesis—that MNPI about one bank would affect similar banks generally. However, the positive trendline indicates that the magnitude of those gains completely decoupled from semantic similarity rank: lower-similarity peers outperformed higher-similarity peers. This is the sector-wide contagion failure mode: the merger signaled regulatory approval of large-bank consolidation broadly, lifting all regional bank peers regardless of their textual similarity to SunTrust. Because our test evaluates the strong form of the hypothesis— whether similarity rank predicts return magnitude, which is what the SEC needs to identify specific trading targets—this case is classified as contradicting, even though all peers moved in the predicted direction.

BB&T/SunTrust similarity rank versus abnormal return

Figure 28. BB&T / SunTrust Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
MTB0.0311143602741742260.041091210736052076Increased
RF0.0064891733912550330.016466023853132884Increased
FITB0.0022457200984048340.012222570560282685Increased
CFG0.0307070606451996170.04068391110707747Increased
KEY0.035173165783291040.04515001624516889Increased
PNC0.00187590895743485290.011852759419312704Increased
PNFP0.026992110799930780.03696896126180863Increased
FHN0.027479945946387160.037456796408265006Increased
HBAN0.04334829587265290.053325146334530754Increased

Table 63. Abnormal Returns SunTrust

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1RF0.910.00648917339125500.0164660238531328Increased0.014984081706350849
2KEY0.880.03517316578329100.0451500162451688Increased0.03973201429574855
3FITB0.870.00224572009840480.0122225705602826Increased0.010633636387445862
4CFG0.860.03070706064519960.0406839111070774Increased0.03498816355208656
5PNC0.850.00187590895743480.0118527594193127Increased0.010074845506415795
7HBAN0.780.04334829587265290.0533251463345307Increased0.041593614140933946
8MTB0.740.03111436027417420.0410912107360520Increased0.03040749594467848

Table 64. Final Analysis SunTrust

KeyCorp / First Niagara

KeyCorp/First Niagara is classified as strongly supporting. The top-ranked firm (CFG) shows a positive AR while all lower-ranked firms show negative reactions, producing a clear negative trendline that aligns with the shadow trading hypothesis. This is the cleanest supporting case in the Finance subsample.

KeyCorp / First Niagara Similarity vs. Abnormal Returns

Figure 29. KeyCorp / First Niagara Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
BBT-0.03606357708124112-0.02173759961246282Decreased
CFG0.0078805700141190510.02220654748289735Increased
WBS-0.027267811913015836-0.012941834444237537Decreased
FNB-0.023912975220883568-0.009586997752105269Decreased
CBU-0.0187772386177642-0.0044512611489859Decreased
VLY-0.018691419065318966-0.004365441596540667Decreased
FULT-0.021865289671713632-0.007539312202935333Decreased
MTB-0.025768358038338268-0.011442380569559969Decreased

Table 65. Abnormal Returns First Niagara

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1CFG0.920.00788057001411900.0222065474828973Increased0.020430023684265514
2WBS0.85-0.0272678119130158-0.0129418344442375Decreased-0.011000559277601876
3CBU0.79-0.0187772386177642-0.0044512611489859Decreased-0.0035164963076988608
4BBT0.76-0.0360635770812411-0.0217375996124628Decreased-0.01652057570547173
5FULT0.71-0.0218652896717136-0.0075393122029353Decreased-0.005352911664084063

Table 66. Final Analysis First Niagara

People’s / United

People’s/United is classified as contradicting. The trendline is positive and most peers show negative or near-zero ARs. The one strongly positive outlier (MRBK, +4.4% AR) ranks last in similarity, directly inverting the expected pattern.

People’s / United Similarity vs. Abnormal Return

Figure 30. People’s / United Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
NBTB-0.019291550895895102-0.013641987240980431Decreased
WASH-0.007836123577671349-0.0021865599227566777Decreased
BBT-0.0287973922743519-0.02314782861943723Decreased
MRBK0.0386904626226188540.04434002627753353Increased
WNEB-0.00320199926063141370.0024475643942832572Increased
FFIC-0.010700167840185815-0.005050604185271144Decreased
INDB-0.014858426109558339-0.009208862454643668Decreased

Table 67. Abnormal Returns United

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
2INDB0.88-0.0148584261095583-0.0092088624546436Decreased-0.008103798960086369
4WASH0.81-0.0078361235776713-0.0021865599227566Decreased-0.0017711135374328462
5WNEB0.73-0.00320199926063140.0024475643942832Increased0.001786722007826736
7NBTB0.66-0.0192915508958951-0.0136419872409804Decreased-0.009003711579047065
9FFIC0.52-0.0107001678401858-0.0050506041852711Decreased-0.002626314176340972
10MRBK0.420.03869046262261880.0443400262775335Increased0.01862281103656407

Table 68. Final Analysis United

Sterling / Astoria

Sterling/Astoria is classified as supporting. The top-ranked firm (FFIC, +0.79% AR) shows a positive reaction while lower-ranked firms show negative reactions, producing a modest negative trendline. The magnitude of reactions is small throughout, reflecting the limited market impact of a regional bank acquisition.

Sterling / Astoria Similarity vs. Abnormal Returns

Figure 31. Sterling / Astoria Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
VLY-0.004796628053415628-0.0019719103377265033Decreased
CNOB-0.008080841000498168-0.005256123284809043Decreased
FFIC0.0050632614835826870.00788797919927181Increased
FLG-0.007487942790061821-0.004663225074372696Decreased
NFBK-0.006468928281572891-0.003644210565883767Decreased
CFG-0.007839123578677706-0.005014405862988582Decreased
DCOM0.0028168698654325880.005641587581121712Increased
PFS-0.004540318237324739-0.0017156005216356143Decreased
KRNY-0.006557784922778603-0.003733067207089479Decreased

Table 69. Abnormal Returns Astoria

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
2FFIC0.930.00506326148358260.0078879791992718Increased0.007335820655322775
5KRNY0.83-0.0065577849227786-0.0037330672070894Decreased-0.003098445781884202
6NFBK0.82-0.0064689282815728-0.0036442105658837Decreased-0.0029882526640246337
9PFS0.77-0.0045403182373247-0.0017156005216356Decreased-0.001321012401659412
10VLY0.70-0.0047966280534156-0.0019719103377265Decreased-0.0013803372364085498

Table 70. Final Analysis Astoria

WSFS / BENEFICIAL

WSFS/Beneficial is a supporting case. All peers show positive abnormal returns, and the top-ranked firm (UVSP) together with PGC (rank 3) show among the largest positive ARs. The negative trendline holds across the cohort, making this one of the more consistent supporting cases in the Finance subsample. PKBK’s similarity score is recorded as 0.00, which is a transcription error in our Stage 2 output rather than a model judgment; because PKBK ranks last under any plausible score below 0.65, the event’s rank correlation (ρ=+0.20\rho= +0.20) is unchanged by the error.

WSFS/Beneficial: Similarity Rank vs. Abnormal Return

Figure 32. WSFS / Beneficial Similarity vs. Abnormal Returns

TickerRaw ReturnAbnormal ReturnDirection
KRNY0.0073261690069663550.004861293886339441Increased
UVSP0.0090579413298528690.006593066209225955Increased
FRBA0.0179857636110045440.015520888490377631Increased
PKBK0.008583615909959380.006118740789332466Increased
OCFC0.0065994032666816180.004134528146054704Increased
CUBI0.0095165586465041940.00705168352587728Increased
FRBKQ0.0-0.002464875120626914Decreased
FULT0.0171922997379955550.014727424617368642Increased
WSFS-0.08130069889002943-0.08376557401065635Decreased
PGC0.0139015688348273680.011436693714200453Increased

Table 71. Abnormal Returns Beneficial

RankTickerSim. ScoreRaw ReturnAbnormal ReturnDirectionShadow Signal
1UVSP0.940.00905794132985280.0065930662092259Increased0.006197482236672346
3PGC0.870.01390156883482730.0114366937142004Increased0.009949923531354347
4OCFC0.840.00659940326668160.0041345281460547Increased0.0034730036426859475
6KRNY0.700.00732616900696630.0048612938863394Increased0.00340290572043758
7CUBI0.650.00951655864650410.0070516835258772Increased0.00458359429182018
9PKBK0.000.00858361590995930.0061187407893324Increased0.0

Table 72. Final Analysis Beneficial

References

  1. [1]Acquisti, A., Brandimarte, L., and Loewenstein, G. Privacy and human behavior in the age of information. Science, 347(6221):509–514, 2015. doi: 10.1126/science.aaa1465.DOI
  2. [2]Araci, D. Finbert: Financial sentiment analysis with pre-trained language models. arXiv preprint arXiv:1908.10063, 2019.arxiv.org/abs/1908.10063
  3. [3]Augustin, P., Brenner, M., and Subrahmanyam, M. G. Informed options trading prior to takeover announcements: Insider trading? Management Science, 65(12):5697–5720, 2019.DOI
  4. [4]Barr, W. P. The securities and exchange commission is watching you. Wall Street Journal Op-ed, April 2024. URL https://www.hudson.org/legal-affairs/securities-exchange-commission-watching-you-bill-barr. Apr. 15, 2024.
  5. [5]Chen, J. and Sarkar, S. A semantic approach to financial fundamentals. In Chen, C.-C., Huang, H.-H., Takamura, H., and Chen, H.-H. (eds.), Proceedings of the Second Workshop on Financial Technology and Natural Language Processing, pp. 22–26, Kyoto, Japan, 5 January 2020. -. URL https://aclanthology.org/2020.finnlp-1.4/.
  6. [6]de Montjoye, Y.-A., Hidalgo, C. A., Verleysen, M., and Blondel, V. D. Unique in the crowd: The privacy bounds of human mobility. Scientific Reports, 3:1376, 2013. doi: 10.1038/srep01376.DOI
  7. [7]Google DeepMind. Gemini 3.1 Pro. February 2026. URL https://deepmind.google/models/model-cards/gemini-3-1-pro.
  8. [8]Hanley, K. W. and Hoberg, G. Dynamic interpretation of emerging risks in the financial sector. The Review of Financial Studies, 32(12):4543–4603, 2019.
  9. [9]Harshavardhan. Self-anchoring calibration drift in large language models: How multi-turn conversations reshape model confidence, 2026. URL https://arxiv.org/abs/2603.01239.
  10. [10]Hoberg, G. and Phillips, G. Text-based network industries and endogenous product differentiation. Journal of Political Economy, 124(5):1423–1465, 2016. doi: 10.1086/688176. URL https://doi.org/10.1086/688176.
  11. [11]Huang, J., Chen, X., Mishra, S., Zheng, H. S., Yu, A. W., Song, X., and Zhou, D. Large language models cannot self-correct reasoning yet, 2024. URL https://arxiv.org/abs/2310.01798.
  12. [12]Jenkins, J. Insider trading policies: Shadow trading, October 2025. URL https://www.thecorporatecounsel.net/blog/2025/10/insider-trading-policies-shadow-trading.html.
  13. [13]Khot, T., Trivedi, H., Finlayson, M., Fu, Y., Richardson, K., Clark, P., and Sabharwal, A. Decomposed prompting: A modular approach for solving complex tasks, 2023. URL https://arxiv.org/abs/2210.02406.
  14. [14]Koval, R., Andrews, N., and Yan, X. Learning to compare financial reports for financial forecasting. In Findings of the Association for Computational Linguistics: EACL 2024, pp. 500–512, St. Julian’s, Malta, 2024. Association for Computational Linguistics. URL https://aclanthology.org/2024.findings-eacl.34/.
  15. [15]Li, J., Chen, J., Ren, R., Cheng, X., Zhao, W. X., Nie, J.-Y., and Wen, J.-R. The dawn after the dark: An empirical study on factuality hallucination in large language models, 2024. URL https://arxiv.org/abs/2401.03205.
  16. [16]Mehta, M. N., Reeb, D. M., and Zhao, W. Shadow trading. The Accounting Review, 96(4):367–404, 2021. URL https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3689154.
  17. [17]New Civil Liberties Alliance. Complaint, Davidson, Restivo, and NCPPR v. Atkins, no. 6:24-cv-00197 (w.d. tex. apr. 16, 2024), April 2024.
  18. [18]Peirce, H. This CAT is a dangerous dog. RealClearPolicy, October 2019. URL https://www.realclearpolicy.com/articles/2019/10/09/this_cat_is_a_dangerous_dog_111285.html.
  19. [19]Plaintiffs. Memorandum in support of renewed motion for a preliminary injunction and stay pursuant to 5 U.S.C. § 705, Davidson v. Atkins, No. 6:24-cv-00197-ADA-DTG. W.D. Tex., January 2026. Filed Jan. 30, 2026.
  20. [20]Securities and Exchange Commission. SEC Approves New Rule Requiring Consolidated Audit Trail, Exchange Act Release No. 34-67457. Technical report, U.S. Securities and Exchange Commission, July 2012. 77 Fed. Reg. 45,722 (Aug. 1, 2012).
  21. [21]Solove, D. J. A taxonomy of privacy. University of Pennsylvania Law Review, 154(3):477–560, 2006. URL https://ssrn.com/abstract=667622.DOI
  22. [23]Stenfors, A., Guo, T., Li, B., Hewage, K., Mere, P., and Chen, F. Shadow trading detection: A graph-based surveillance approach. Finance Research Letters, 86: 108524, 2025. ISSN 1544-6123. doi: https://doi.org/10.1016/j.frl.2025.108524. URL https://www.sciencedirect.com/science/article/pii/S1544612325017787.
  23. [24]Supreme Court of the United States. United States v. O’Hagan, 521 U.S. 642, 1997.
  24. [25]Supreme Court of the United States. City of Los Angeles v. Patel. 576 U.S. 409, 2015.
  25. [26]Supreme Court of the United States. Carpenter v. United States, 585 U.S. 296, 2018a.
  26. [27]Supreme Court of the United States. Sessions v. Dimaya, 584 U.S. 148, 2018b. Gorsuch, J., concurring.
  27. [28]Supreme Court of the United States. Chatrie v. United States, No. 25-112. 609 U.S. __ (2026), June 2026. Decided June 29, 2026 (Kagan, J.); vacated and remanded, 6–3, reviewing United States v. Chatrie, 136 F.4th 100 (4th Cir. 2025) (en banc).
  28. [29]The Brattle Group. Shadow trading: Economic evidence and potential implications from recent SEC insider trading cases. Brattle Group White Paper, November 2025. URL https://www.brattle.com/wp-content/uploads/2025/11/Shadow-Trading-Economic-Evidence-and-Potential-Implications-from-Recent-SEC-Insider-Trading-Cases.pdf.
  29. [30]United States District Court, N.D. Cal. SEC v. Panuwat, 702 F. Supp. 3d 883, 2023. No. 21-cv-06322-WHO.
  30. [31]U.S. Securities and Exchange Commission. SEC v. Bechtolsheim. Settled May 2024; civil penalty of $923,740 without admission of liability, May 2024.
  31. [32]U.S. Securities and Exchange Commission. Order approving an amendment to the national market system plan governing the consolidated audit trail, as modified by amendment nos. 1 and 2 and by the commission, regarding customer and account information. Exchange Act Release No. 34-104586, 91 Fed. Reg. 2,164, January 2026. Approved Jan. 13, 2026.
  32. [33]Wang, P., Li, L., Chen, L., Cai, Z., Zhu, D., Lin, B., Cao, Y., Kong, L., Liu, Q., Liu, T., and Sui, Z. Large language models are not fair evaluators. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 9440–9450. Association for Computational Linguistics, 2024. URL https://aclanthology.org/2024.acl-long.511/.arxiv.org/abs/2201.11903
  33. [34]Wei, J., Wang, X., Schuurmans, D., Bosma, M., Ichter, B., Xia, F., Chi, E., Le, Q., and Zhou, D. Chain-of-thought prompting elicits reasoning in large language models, 2023. URL https://arxiv.org/abs/2201.1903.
  34. [35]Wu, S., Irsoy, O., Lu, S., Dabravolski, V., Dredze, M., Gehrmann, S., Kambadur, P., Rosenberg, D., and Mann, G. BloombergGPT: A large language model for finance. arXiv preprint arXiv:2303.17564, 2023.arxiv.org/abs/2303.17564

Paper details

vlnc:W_zz8xeewq
Version
v1 · latest
Published
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

Not recorded for this paper

View statement
Looking for related papers…
Contents