BaselineWatch APEX Weekly — institutional disclosure intelligence
In the current macroeconomic environment—a period marked by institutional friction rather than sudden shocks—financial markets are mispricing policy risk. This mispricing is not rooted in information scarcity but rather in the market’s attention architecture. Investors focus on headlines, earnings calls, and guidance updates, but the true policy transmission mechanism has shifted elsewhere: into routine 10-Q filings. Specifically, the linguistic substrate of risk-factor updates has emerged as the first observable signal of policy-related uncertainty.
The mechanism is structural. Unlike 10-K filings, which restate the entire risk landscape annually, 10-Qs are obligated to update risk factors only for material changes. Yet our data shows that subtle but significant shifts in language—more modal caution (“may,” “could”), greater specificity in conditional statements, and rising uncertainty density—precede material events that force repricing. This phenomenon is not speculative; it is validated by BaselineWatch’s proprietary Flagship signal cohort, which tracks 10-Qs exhibiting concurrent rises in negative and uncertainty language, bracketed by a material 8-K within a 90-day window. The cohort delivers a mean 60-day return of +6.72% with a t-statistic of 11.95 (n=850), and the effect survives standard Fama-French 5-factor controls, preserving a positive unit-beta residual of +1.99% (t = 3.17, p < 0.01).
This signal carries power precisely because it sits in a blind spot of standard portfolio construction practices. Analysts and fund managers are incentivized to chase variables with daily frequency and broad cross-sectional coverage. Disclosure drift, by contrast, is event-driven and sparse, pulling attention away from the slow grind of institutional policy evolution. The result is an inefficiency rooted in attention allocation, not information scarcity. Headlines are efficiently priced; standalone news sentiment strategies fail to produce additive returns (t = ‒8.69, p = 0.49). But the language changes embedded in filings—captured systematically by our NLP-driven analysis—offer a tradable edge.
This edge is particularly visible in sectors most exposed to policy-driven constraints. Utilities and independent power producers are rewriting risk factors to account for siting and interconnection challenges even as their revenue narratives remain stable. Insurers adjust capital and tax language ahead of actuarial or regulatory shifts. Logistics firms and postal-adjacent incumbents caveat service obligations in response to evolving compliance standards while operational volumes stay steady. In each case, management’s first move is linguistic, reflecting a recalibration of how they perceive timing, scope, and potential downside. The second move, if it arrives, is an 8-K—an enforcement notice, permit delay, rate-setting outcome, or other event that confirms the linguistic signal.
This phenomenon is not random noise. It is rooted in the political economy of policy transmission. Lawmakers increasingly pass skeletal frameworks that delegate detail to agencies. Agencies, in turn, issue rules iteratively, subject to court challenges and industry feedback. States modify or countermand federal templates, creating a fragmented compliance landscape. Each step introduces ambiguity—license timing, capital treatment, tax interpretation—that must be translated into disclosure language. That translation inevitably increases hedging and specificity, even if guidance and operational metrics appear steady. Our Flagship signal cohort formalizes this translation into an actionable framework: watch for concurrent rises in negative and uncertainty language in 10-Q risk factors, then monitor for a confirming event within the next quarter.
The implications are clear. First, policy drift will persist as a macro force because its institutional roots—delegation, fragmentation, and judicial review—are structural and long-lasting. Second, the repricing mechanism favors those who monitor filings rather than headlines. Markets underreact to the initial linguistic shift, creating an opportunity for informed investors to act before events force broader attention. Third, this is not a call to treat every uptick in caution as a directional signal. The Flagship signal cohort’s specificity—concurrent linguistic shifts followed by material events—is what makes it tradable. The rest is noise, and our settled verdicts reinforce this: standalone news sentiment strategies fail, disclosure-elevation artifacts are nonviable, and earnings-call language programs do not deliver alpha.
In sum, the current cycle is defined by institutional friction and path-dependent policy evolution. Investors who focus on the linguistic signals embedded in 10-Q filings—rather than chasing headlines or earnings-call soundbites—will be better positioned to navigate this environment. The opportunity exists because it is overlooked, and the returns speak for themselves: +6.72% over 60 days, validated across 850 observations, with robust statistical backing. This is the signal the market has yet to see—and the one it cannot afford to ignore.
The broader implications of this signal expand far beyond the immediate returns. For investors, this finding challenges the conventional reliance on traditional market indicators like price volatility and news sentiment. It underscores the importance of rethinking portfolio construction practices to include linguistic drift analysis as a core input. Even sectors seemingly shielded from policy risk—such as technology and consumer discretionary—are not immune to the underlying dynamics of disclosure drift, as regulation around data privacy, labor compliance, and intellectual property enforcement continues to evolve. The institutional grind of policy transmission—with its delegation to agencies, iterative rulemaking, and judicial challenges—means that the policy environment will remain a critical driver of uncertainty and return dispersion across sectors.
For fund managers and institutional investors, the actionable takeaway is clear: the edge lies in detecting linguistic signals early and pairing them with subsequent events. This pattern is repeatable and tradable precisely because standard benchmarks fail to capture it. Attention allocation, not information scarcity, is the root inefficiency. Those who adapt their tools and frameworks to systematically compare disclosure language across time will find themselves ahead of the curve, better equipped to anticipate and act on policy-driven uncertainty.
Importantly, this is not a universal solution. The Flagship signal cohort’s signal distinguishes itself through specificity and robust validation. It is not a blanket call to action for every shift in risk-factor language but rather a targeted framework for identifying concurrent rises in negative and uncertainty language followed by a material event. The rest is noise—a cautionary reminder against overinterpreting isolated linguistic changes in filings without the validating event sequence.
THEME — Quiet Policy Risk Is Migrating Into Routine 10‑Qs
The theme in one sentence: The market is focused on headlines and hearings, but the real policy transmission mechanism is showing up in routine 10‑Q risk‑factor updates — a steady creep of uncertainty language that investors ignore until a material 8‑K follows, when it is already tradable.
Thesis
Policy risk in 2026 is less about sudden shocks and more about a slow institutional grind — legislative ambiguity, rulemaking backlog, and enforcement variability filtering into corporate disclosure. That grind does not reliably surface in guidance or newsflow; it registers first in the linguistic substrate of quarterly filings. Our disclosure‑language data repeatedly shows that when a company’s 10‑Q exhibits a concurrent rise in negative and uncertainty language and is followed or bracketed by a material event, the post‑filing return path is not neutral — it is systematically different from the baseline. That is the core of our crown‑jewel cohort: 10‑Qs with rising negative and uncertainty language where a material 8‑K lands within a 90‑day window. The validation is locked: the cohort’s mean 60‑day return is +6.72% with a t‑statistic of 11.95 (n=850), and the effect survives standard factor controls in our secondary specification. You do not need earnings calls or paid newsfeeds to see this; it is in the filings.
What other outlets are missing is the locus of that drift. Commentary gravitates to technology multiples and the latest macro print. But the disclosure language is shifting across policy‑exposed corners of the real economy where rules and permits matter more than demand elasticity. Utilities and independent power producers rewrite risk factors around licensing and grid policy while guidance stays uncontroversial. Insurers adjust capital and tax language before the first rate quote moves. Logistics and postal‑adjacent incumbents caveat service obligations while volumes still look steady. In each case, management’s first move is linguistic: more modal caution (“may/could”) and more specificity in risk narratives. The second move — if and when it comes — is an 8‑K. Our data says the time to pay attention is the first move.
Evidence and mechanism
The mechanism is structural. Quarterly reports do not require risk‑factor updates unless there are material changes. When risk sections shift anyway — rising uncertainty density alongside an uptick in negative terminology — it is management acknowledging a policy vector that is hard to price and easy to litigate. The market underweights these shifts because they are text, not numbers, and because news‑sentiment overlays add little signal on their own. We tested that proposition explicitly. Our news layer is conditioning‑only; standalone news‑alpha was rejected in replication and in our internal distribution tests. By contrast, disclosure‑language drift validated out‑of‑sample and remains operational. The definable, tradable cohort is narrow: 10‑Qs that show rising negative and uncertainty language with a material 8‑K inside a 90‑day band. Within that, the post‑filing path is statistically distinct from the universe (mean 60‑day +6.72%, t=11.95; locked result). The lesson is not that “bad words” are bearish; it is that change in risk signaling, when paired with a proximate event, carries information the market processes slowly.
This is why the current moment matters. The live tape is saturated with policy noise — energy licensing fights, tax reconsolidation proposals, sector bills that advance and stall. Lobbying disclosures frame the front end of the pipeline, but it is the 10‑Q that records how a company internalizes the resulting uncertainty. We do not need to predict which bill passes. We need to detect when management stops treating the risk as boilerplate and starts writing it into the quarterly record. That turn is measurable in our corpus and precedes the market’s repricing far more often than news‑flow heuristics imply.
Angles to commission (deep dives)
1) Permits, prudence, and the grid — utilities and independent power
The question: are utilities and IPPs rewriting quarterly risk narratives around nuclear licensing, interconnection queues, and rate‑case cadence before capital plans move? We will scan sector 10‑Qs for rising uncertainty and negative‑language density in risk sections and MD&A, focusing on firms with recent regulatory touchpoints and near‑term 8‑K activity (permits, regulatory decisions, or capital structure notices). The objective is to isolate where text moved first and whether a material event followed within the crown‑jewel window. Expect a concentration in utilities with heavy capex programs and merchant generators juggling policy‑driven project risk.
2) Statutory capital and the tax turn — insurers and balance‑sheet businesses
Insurers and other balance‑sheet businesses are canaries for policy shifts: they file quarterly, are tightly regulated, and must surface capital and tax‑base risks promptly. We will examine 10‑Qs for drift in risk‑factor specificity around statutory capital requirements, deferred tax assets, and prospective changes to loss‑reserve treatments. Where those linguistic shifts sit next to a subsequent 8‑K (capital actions, reserve updates, or regulatory correspondence), we will test for membership in the crown‑jewel cohort and map any sector‑level pattern. The goal is to catch the policy turn when it is still “words,” not capital costs.
3) State capacity and service obligations — logistics and postal‑adjacent incumbents
A different channel of policy risk runs through service mandates and federal partnerships. Postal‑adjacent and parcel integrators sit where public obligations and private economics meet. We will track 10‑Q risk‑factor updates that elevate dependency on federal service levels, contract structures, or regulatory definitions of universal service. If those updates are soon followed by 8‑K disclosures (contract amendments, compliance notices, or pricing changes), they drop into our high‑signal cohort. The brief is to quantify where the language moved first and whether the subsequent events fit the same text‑before‑event pattern.
Where the evidence will come from
We will not rely on headline counts or sector PR to anchor this. The evidence sits in specific, tradable cohorts within our corpus:
– Crown‑jewel cohort (operational): 10‑Qs where negative and uncertainty language rises versus the prior comparable filing and a material 8‑K lands within 90 days. This is the backbone. The validation is locked at +6.72% mean 60‑day return with t=11.95 (n=850); factor‑controlled variants corroborate the effect in secondary tests. Our scans will shortlist members in the current filing season across utilities/power, insurers and balance‑sheet financials, and logistics/postal‑adjacent firms.
– Rate‑sensitive mid‑caps with regulatory cadence: companies whose economics hinge on permits, statutory capital, or regulated pricing. The hypothesis is not sector beta but disclosure behavior: a measurable turn in risk‑factor wording in 10‑Qs ahead of rate cases, capital notices, or service‑obligation changes recorded as 8‑Ks.
– Event‑bracketed drift in the real economy: mid‑cycle industrials and infrastructure names that face policy‑driven input or tariff changes. Where filings show the uncertainty‑plus‑negative language shift and are bracketed by supplier notices, tariff updates, or board actions in 8‑Ks, they meet the same tradable profile.
We will not use news‑sentiment overlays to select targets; those are conditioning‑only in our framework and add little on their own. Nor will we rely on earnings‑call tone — a program we closed after pre‑registered variants failed to deliver a robust signal. The core filter is filing language drift plus proximate event. Within that, we are sector‑agnostic except as a lens for investigative context.
Implications
For allocators: this is a monitoring problem, not a forecasting one. The signal is not about predicting which bill passes; it is about catching when policy uncertainty becomes material enough that counsel insists on rewriting the quarterly record. In our data, when that linguistic turn coincides with a material event in the same window, the market takes time to price it. That lag is investable in our crown‑jewel setup and robust to standard factor controls in secondary analysis. The policy regime’s slow grind thus argues for a workflow change: promote quarterly risk‑factor drift from a compliance checkbox to a standing, cross‑sector alert.
For issuers: the gap between “we may be affected” and “we are likely to be affected” is linguistically small but market‑relevant. Boilerplate persistence and word‑count growth alone do not carry the signal; change does. Firms that treat risk sections as static templates miss an opportunity to control the narrative early. Conversely, those that transparently update risk sections before events often find the market credibly interprets the change only after the follow‑on 8‑K — when optionality is lower.
For policymakers: the disclosure record is quietly aggregating the costs of rulemaking delays and enforcement variability. If the language is drifting before the rules are finalized and before companies alter capital plans, the filings are an early ledger of regulatory friction. It is visible, quantifiable, and, in our corpus, linked to measurable post‑filing return paths in event‑bracketed quarters. That should inform both consultation timelines and communication clarity.
The editorial North Star for the week
We will center the issue on where policy risk actually enters markets: not at the podium or in the tape, but in the legal prose of 10‑Qs. We will map the drift, select crown‑jewel members in the current season across utilities, insurers, and logistics, and report the event brackets. The conclusion we expect to test is simple: the first tell is the change in words; the trade, if any, belongs to the quarter when words and events finally line up.
Quiet policy risk is migrating into routine 10‑Qs because the policy apparatus itself has shifted from discrete, date‑stamped acts to a slow institutional grind. The channels are familiar—legislation, rulemaking, enforcement, and adjudication—but their cadence and interaction have changed in ways that disclosure language registers before markets do. Agencies face longer comment cycles and more remands; states pull in different directions; enforcement posture varies with leadership and docket pressure; compliance deadlines arrive in phases with carve‑outs and conditional safe harbors. None of that produces a clean headline in real time. It does, however, force management teams to revise how they describe risk and uncertainty even when they keep guidance steady and avoid splashy announcements. That is why the first observable movement is linguistic, not financial: more modal caution (“may,” “could”), more hedging around timing and scope, and more specific caveats inserted into risk factors that 10‑Qs are not required to update unless something material has changed.
This is not a claim about a single law or a single agency; it is a claim about institutional friction as the dominant macro force in 2026. Monetary conditions have cooled from the violent regime shifts of the last few years, and the news cycle has normalized around a predictable schedule of data releases and hearings. In that environment, investors gravitate to what is easy to timestamp and backtest—prints, press releases, conference‑call soundbites—while underweighting disclosures whose signal is embedded in changes relative to a prior baseline. The result is a pricing gap around policy transmission. Our data shows that when policy risk begins to bite, companies first change how they talk about it in the filings, and only later does an 8‑K or other event force a repricing.
Two empirical points anchor this view. First, headlines themselves do not carry tradable power for this use case. In a settled verdict on our news layer, a standalone news‑sentiment strategy failed to produce additive returns (t = −0.69, p = 0.49), even before cluster adjustments. The implication is not that information is useless but that the market digests headline tone efficiently; there is no residual edge in chasing it. Second, disclosure drift does carry power when it coincides with real events. Our “crown‑jewel” cohort—10‑Qs that show a concurrent rise in negative and uncertainty language, bracketed by a material 8‑K within a 90‑day window—delivers a mean 60‑day return of +6.72% with a t‑statistic of 11.95 (p < 0.001; n = 850). A secondary factor‑controlled specification preserves a positive unit‑beta residual of +1.99% (t = 3.17, p < 0.01). These figures are locked in our canon and speak to the mechanism: markets underreact to the first move (language), then reprice when the second move (event) lands. You do not need paid feeds or privileged access to observe this sequence; it is literally in the public filings.
Why now? Because the policy pipeline has become path‑dependent and iterative. Lawmakers increasingly pass skeletal frameworks that delegate specificity to agencies. Agencies, in turn, issue rules in tranches, subject to court challenge and revision. Industry groups negotiate compliance via comment letters and interim guidance. States copy, modify, or countermand the federal template. Enforcement calibrates quietly as caseloads and leadership priorities shift. Each step introduces ambiguity that is real for operators—license timing, capital treatment, tax interpretation, data‑handling protocols, reporting thresholds—and that ambiguity must surface somewhere. It rarely hits forward guidance first; it lands in Item 1A risk factors and the MD&A, in the verbs and qualifiers managers choose when they explain what could go wrong.
The sectors where this shows up earliest are not the headline leaders of the last cycle. They are the policy‑exposed corners of the real economy where permits, licenses, capital rules, and service obligations dominate. Utilities and independent power producers are rewriting risk factors around siting and interconnection even as their near‑term revenue narratives remain steady. Insurers adjust capital and tax language ahead of any move in pricing tables. Logistics firms and postal‑adjacent incumbents caveat service levels and contractual performance around evolving regulatory standards while volumes still look acceptable. In each case, management’s first signal is linguistic: more “may” and “could,” fewer hard commitments, and expanded specificity about what depends on what. The second signal, if it arrives, is an 8‑K—an enforcement notice, a permit delay, a rate‑setting outcome, a capital action—that confirms the drift that was already on the page.
Markets have not fully priced this because the attention architecture points elsewhere. Screens and factor models are built to digest levels and surprises, not changes in text. Portfolio construction practices privilege variables with daily frequency and cross‑sectional breadth; disclosure drift is event‑driven and sparse by design. Coverage incentives pull analysts toward earnings calls and management meetings, not toward line‑by‑line comparisons of risk sections across quarters. And policy fatigue is real: after a decade of debates over tariffs, pandemics, and bank capital, investors want to believe the regime has stabilized. That desire for stability biases the tape toward over‑weighting the visible and under‑weighting the durable.
The institutional constraints on disclosure amplify the signal. Unlike 10‑Ks, 10‑Qs are not obligated to restate the entire risk universe; they must update only for material changes. When a 10‑Q’s risk factors get longer, more detailed, and more hedged, the default presumption should be that something material has changed—even if guidance holds and the call script stays bland. That is precisely the condition our crown‑jewel cohort formalizes. The statistical outcome is not ambiguous: +6.72% average 60‑day returns (t = 11.95, p < 0.001) across 850 observations, with a positive residual in a unit‑beta FF5 specification (t = 3.17, p < 0.01). The mechanism is straightforward: language moves first, events catch up, and the repricing happens in the window ordinary investors can trade.
Importantly, the null on news alpha (t = −0.69, p = 0.49) resolves a common objection. If news is efficiently priced and yet filings continue to carry incremental information via their changes in wording, then the locus of underreaction is not information scarcity; it is attention allocation. The market reads headlines as standalone objects. It does not systematically compare today’s risk language to last quarter’s, and it is not trained to interpret a higher density of uncertainty terms as a state change. That is a tractable inefficiency.
A political‑economy reading reinforces this. Policy is being made in slow motion through guidance updates, enforcement memos, and case‑by‑case adjudications. Legislatures outsource detail to agencies; courts force agencies to defend detail at length; agencies rely on iterative compliance pathways to keep industry at the table. The burden of translation falls on corporate counsel and CFOs: update the risk section to reflect what your lawyers now believe is possible or likely. That translation inevitably increases the density of hedging and the breadth of contingencies, even if the CFO would rather say “no change.” The disclosure signal’s power is that it quantifies this translation consistently across thousands of issuers and decades. When negative language and uncertainty language both rise in tandem—and a material event lands within the next quarter—the returns are not random noise; they are statistically and economically meaningful under standard controls (p < 0.001 on the headline result; p < 0.01 on the residual specification).
What does this mean for the next leg of the cycle? First, policy drift will persist as a macro driver because the institutional roots—delegation, fragmentation, and judicial review—are not going away. Second, the repricing mechanism will continue to favor those who watch filings rather than headlines. As agencies calibrate enforcement and as state‑federal divergence widens, expect more companies to pre‑signal in Item 1A and MD&A that the ground is shifting. Some of those companies will then deliver the confirming 8‑K that makes the language tradable. The opportunity exists precisely because it sits in a blind spot of the standard toolkit.
Third, this is not a call to treat every uptick in caution as a sell signal or a buy signal. The signal that clears our battery is specific: concurrent rises in negative and uncertainty language in the 10‑Q, with a material event in proximity. The rest is noise, and the nulls in our ledger are instructive. Attempts to extract stand‑alone alpha from news sentiment failed (t = −0.69, p = 0.49). Attempts to carry over earnings‑call language features into a tradable signal were adjudicated null as well. The implication is narrow but powerful: focus on what managers must write down when policy risk increases, not on what they say elsewhere or what the headlines infer.
Finally, the process edge is practical. This is public data, timestamped by EDGAR, and tradable without exotic plumbing. It does not require anticipation of legislative votes or inside knowledge of agency calendars. It requires attention to the linguistic substrate of disclosure and respect for the fact that in a slow‑policy regime, words move before actions do. The crown‑jewel result (t = 11.95, p < 0.001; n = 850) is not a backtest curiosity; it is an empirical statement about how policy risk transmits into prices when the institutional machine grinds rather than jolts. The market does not price that first move consistently. That is the gap this report is designed to close—by starting where the signal actually starts: the routine 10‑Q that suddenly says “may” and “could” a lot more than it used to.
Deep Dive A — Quiet policy risk is migrating into routine 10‑Qs
The policy channel that matters in 2026 is not the spectacle of hearings or the spasm of headline risk. It is the slow, administrative grind that filters into the text of quarterly filings. In our data, the investable inflection is the moment a company’s 10‑Q shows a concurrent rise in negative and uncertainty language and that filing is followed or bracketed by a material 8‑K within a 90‑day window. That cohort — the crown‑jewel we locked in April — delivers a +6.72% mean 60‑day return with a t‑statistic of 11.95 (n=850), and the effect survives standard factor controls in our secondary specification. You do not need earnings calls or premium news feeds to see it; it is sitting in Item 1A and MD&A.
This week’s deep dive traces how that drift is showing up where policy meets operating permits, capital requirements, and service obligations — not in momentum names, but in policy‑exposed corners of the real economy. Utilities and independent power producers are rewriting risk factors around licensing and grid policy while guidance stays even. Insurers are adjusting capital and tax language ahead of the first premium change. Logistics and postal‑adjacent incumbents are caveating service and network constraints before volumes visibly roll. In each case, management’s first move is linguistic: more modal caution (“may,” “could,” “might”) coupled with more specificity in the risk narrative. The second move — if and when it comes — is a discrete 8‑K. Our data says the time to pay attention is the first move.
Utilities and independent power producers: the interconnection and permitting channel
Quarterly reports do not require risk‑factor updates unless there are material changes. When a utility still expands or sharpens Item 1A mid‑year, it is a tell. The language has migrated from generic “changes in regulation may affect our business” to concrete mechanics that bind operations to administrative calendars: “progress through interconnection queues,” “timing of receipt of required permits,” “prudence reviews and cost recovery determinations,” “compliance with evolving NERC standards,” and “implementation risks tied to FERC rulemakings.” We also see filings naming dependencies that were previously implied: “state commission approval of recovery mechanisms,” “clarity around eligibility and transferability of energy‑related tax credits,” “grid‑upgrade cost allocation,” and “environmental rulemaking affecting generation mix.”
Two tonal shifts matter for investors. First, modal creep: sections that formerly read “may adversely affect” begin to carry stacked conditionals that lengthen the chain of things that have to go right (“timely receipt of approvals, availability of key equipment, and favorable determinations in rate proceedings”). Second, specificity: broad “environmental regulation” becomes “compliance with federal performance standards and state implementation plans” or “adherence to interim guidance pending final rules.” Those changes are not semantic ornament; they reflect a manager who has moved from hypothetical geopolitics to a live docket of filings, hearings, and engineering lead times. The operational corollary shows up in MD&A: “supply‑chain duration for transformers and switchgear,” “construction sequencing adjusted to reflect interconnection timing,” “capital plan rephased pending cost‑recovery clarity.”
In the 8‑K stream, the anchors are familiar: an Item 2.06 (material impairment) tied to a project that missed a deadline aligned to permitting; an Item 8.01 noting a commission order on cost recovery; an Item 1.01 disclosing amended vendor terms for grid‑critical equipment. Our crown‑jewel configuration is built for exactly this: the language turns, and an event lands within the same 90‑day window. The market habit is to trade the order; our data argues you harvest the drift that precedes it.
Insurers: the capital and tax channel
Property‑casualty and multiline insurers are living in a regime where regulatory change shows up first as capital math and tax interpretive risk, not as an observable premium line. When risk‑factor sections shift mid‑year, the edits are revealing. We see language that moves from “our results may be affected by regulatory action” to “changes in statutory capital requirements and risk‑based capital (RBC) factors,” “availability and cost of reinsurance under evolving treaty terms,” and “uncertainty regarding the timing and substance of tax guidance affecting the treatment of loss reserves or credit monetization.”
Filing text increasingly links underwriting posture to administrative dependencies: “timing of approvals from state departments of insurance,” “adverse determinations in rate‑filing proceedings,” “changes to catastrophe modeling assumptions required by regulators,” and “dividends from regulated insurance subsidiaries subject to prior approval.” Insurers also splice in more precise caveats around investment portfolios: “asset‑liability management under volatility and spread risk,” “valuation of alternative investments under evolving disclosure standards,” and “counterparty risk from financial institutions subject to additional capital or liquidity rules.”
The MD&A mirror is subtle but critical: “reinsurance placements completed at higher attachment points,” “increased ceded premium reflecting tightened retrocession markets,” “reserve reviews under updated assumptions,” and “capital deployment adjusted in anticipation of regulatory changes.” The 8‑Ks that bookend these shifts are often “Other Events” around catastrophe estimates, reserve strengthening disclosures, or announcements of rate‑filing outcomes in key states. The point is not that every insurer is changing language each quarter — many 10‑Qs explicitly state “no material changes” — but that when Item 1A does change, it is more likely to be about the machinery of policy than the weather. Our result says that conjunction — rising negative and uncertainty language, then a material event — is where the tradable path begins.
Logistics and postal‑adjacent incumbents: the service‑obligation channel
For the large integrated carriers and postal‑adjacent incumbents, policy risk is not an abstraction; it is the architecture of obligations they must meet under labor agreements, transportation security rules, and government‑contract terms. Over the last few quarters, 10‑Qs have tightened the language around these dependencies. Where the risk section used to group “regulatory and legal risks,” we now see items that specifically tie network economics to administrative processes: “compliance with transportation security directives,” “limitations on delivery exceptions,” “service‑standard adjustments under government contracts,” and “approval processes affecting rate changes.”
The modal language also migrates. “Could affect” becomes “will constrain network flexibility if implemented as proposed,” or “may require additional investments in fleet and facility technology to comply with evolving security standards.” We see more precise caveats around infrastructure: “availability and lead times for aircraft and vehicle parts,” “constraints in acquiring airworthiness certifications for modified equipment,” and “permitting timelines for facility upgrades.” In MD&A, the operational translation shows up as “air capacity rebalanced to reflect revised service commitments,” “linehaul operations consolidated to meet cost targets,” and “pricing changes paced to contractual and regulatory windows.”
Events that match the text are often distinct from earnings: an 8‑K disclosing contract amendments with government entities, a notice of workforce actions tied to network redesign, or a disclosure of a cybersecurity incident whose remediation plan locks in additional compliance spend. The filings tell you a quarter ahead that network geometry is being dictated by rules that are still in motion.
Why language first — and why not headlines?
Settled testing on our side shows that news sentiment alone does not add incremental alpha to filing‑defined cohorts; we use it for conditioning only. That matters because the visible part of the policy story is the least predictive. Headlines chronicle the political theater. The economic transmission mechanism lives in forms, orders, and guidance — signals the market’s default inbox doesn’t prioritize. Item 1A is the place where management is obligated to encode those constraints with enough specificity to stand up to a hindsight test. When that encoding accelerates — more uncertainty and more negative framing, grounded in concrete dependencies — the repricing that follows is not random drift. Our crown‑jewel cohort quantifies it: +6.72% over the 60 trading days after the filing, t=11.95 (n=850), robust under standard factor controls in our secondary spec.
What to watch — and how to act
For utilities and independent power producers, read for interconnection, permitting, and recovery mechanics: the density of “timing of approvals,” “prudence reviews,” and “eligibility for credit monetization” clauses is the give‑away. Note any pairing in the subsequent 90 days with 8‑Ks that record commission orders, cost‑recovery decisions, or capital‑plan rephasing. For insurers, isolate RBC and tax‑interpretation language, the specifics of reinsurance market access, and the wording around reserve reviews; bracket those filings with 8‑Ks that tighten loss estimates or disclose rate‑filing outcomes. For logistics and postal‑adjacent carriers, track the shift from generic compliance to named obligations and standards, and then watch for contract amendments, network redesign actions, or security‑compliance 8‑Ks. In all three cases, the operative signal is the linguistic substrate, not the subsequent press lines.
This is, by design, a boring channel. It rewards teams that are willing to read the filing text in its institutional context and to accept that policy risk in 2026 is cumulative rather than episodic. The market will continue to reward the loud moments — a rule finalized, a hearing held, a politician quoted — and then overreact to the first concrete event that translates policy into cash‑flow consequences. Our evidence says the repricing starts earlier, when management begins to write to that future. If you are underwriting capital allocation against policy exposure, that is where your edge sits: in the sentence that lengthens the chain of dependencies, in the paragraph that names the docket, and in the pattern where that change in language is followed by a material 8‑K inside the 90‑day window. That is the moment when the policy pipe stops being theoretical and starts dictating the operating model — and when the 60‑day path, on average, ceases to be neutral.
Quiet policy risk in insurance looks nothing like the grid-and-permitting story in utilities, and that divergence is the point. The same macro force—a slow, uneven policy process that leaves companies operating in partial shadow—drives two distinct linguistic signatures. In utilities and independent power, risk sections swell with procedural nouns and docket citations; the voice grows more legalistic as management documents process risk (permits, approvals, interconnection queues). In insurance and diversified financials, the shift is narrower and more modal: more hedging around capital adequacy and tax position, more specificity in reinsurance and exposure concentration, and more conditional phrasing around rate actions and reserving. Both reflect the same policy grind. Only one is written like an operations memo; the other reads like a capital note.
Start with the mechanism. Quarterly reports only require risk‑factor updates when management believes material changes have occurred. For most incumbents, risk sections are a template. A deviation, then, is not a stylistic accident—it is management acknowledging that a policy vector has moved enough to warrant canonizing that movement in a filing. Our data shows that when this deviation takes the form of a concurrent rise in negative and uncertainty language in a 10‑Q, and a material 8‑K lands within the surrounding 90‑day window, the post‑filing return path is not baseline. That cohort—locked as Flagship Disclosure Drift Signal—delivers a +6.72% mean 60‑day return with a t‑statistic of 11.95 (p<0.001; n=850). A secondary FF5 specification preserves the effect in unit‑beta form (+1.99%, t=3.17, p<0.01). You do not need earnings calls or premium news feeds to see it; it is in the filings.
The insurer signature differs from utilities in three ways, all of which are visible in the language. First, capital frame. Utilities write about process and timing—“approval,” “interconnection,” “compliance” and the like. Insurers write about buffers. The uncertainty terms cluster around solvency and flexibility: “capital adequacy,” “risk‑based capital,” “may be required to contribute additional capital,” “could limit the ability to return capital,” “uncertain impact of tax changes on DTAs.” That is a modal hedging shift on the balance sheet rather than a procedural shift in operations. Second, specificity in contingent protection. In insurance, the risk lens often narrows to reinsurance availability, attachment points, and counterparty credit, with new qualifiers around renewals and retrocession. The words become conditional (“may not be able to renew,” “could require higher collateral,” “uncertain recoverability”), and the uncertainty density rises in sentences that previously were neutral boilerplate. Third, exposure taxonomy. Utilities’ policy sections expand horizontally to cover agencies and dockets; insurers’ sections deepen vertically into exposure categories—catastrophe regions, product lines, litigation classes—often introducing new enumerations and qualifiers without changing headline guidance. Taken together, the insurer 10‑Q reads like management tightening narrative controls around statutory capital and earnings volatility before hard outcomes surface.
Why does this matter for investors? Because the market digests these two signatures differently. Process‑heavy drift in utilities is easy to dismiss as regulatory choreography until a project gets delayed; capital‑modal drift in insurers is easier to miss entirely because it hides in specialist vocabulary and small changes to qualifiers. Yet the trading reality we validate is the same: the “first move” is linguistic; the “second move” is an event. In the crown‑jewel cohort, that event is a material 8‑K within 90 days of the 10‑Q with rising negative and uncertainty language, and the mean 60‑day return from the filing date is +6.72% (t=11.95, p<0.001). The persistence of the effect under a standard Fama‑French five‑factor control in our secondary specification (unit‑beta +1.99%, t=3.17, p<0.01) indicates this is not just a proxy for risk premia. The policy grind is being priced—but not immediately, and not symmetrically across language forms.
Consider how these signatures assemble. In utilities, the linguistic move is from general to procedural: risk factors that once said “may face regulatory delays” become “timelines for interconnection are uncertain and may be extended pending System Impact Study completion,” with agency names and specific queue stages spelled out. Negative terms cluster around “delay,” “adverse,” and “compliance,” but the lift in uncertainty comes from modal verbs and temporal caveats (“may,” “could,” “uncertain,” “timing”). In insurers, the move is from high‑level to capital‑calibrated: “may be affected by regulatory changes” becomes “changes in capital standards may require us to hold additional risk‑based capital,” and “we may not be able to fully utilize deferred tax assets in certain jurisdictions.” Negative terms tilt toward “impairment,” “adverse development,” “loss,” and “deterioration,” while uncertainty grows in qualifying phrases around RBC thresholds, reinsurance availability, and rate approvals. The end result is a similar aggregate drift score along our negative+uncertainty dimension, achieved by different lexical pathways.
Crucially, the presence of this drift in insurance does not require an immediate change in pricing, loss ratios, or guidance. In many cases, the filing language moves while the numbers and commentary remain stable. That is the institutional grind at work: a rulemaking timeline slips, an interpretation letter creates ambiguity, a tax proposal introduces optionality—all of it small, none of it conclusive. Management responds by hedging language now and letting the economics catch up. When an 8‑K later records a ratings watch, a reinsurance renewal outcome, an unexpected reserve charge, or a supervisory action, the market treats it as news. Our evidence says the signal was already in the 10‑Q’s words.
What does this mean tactically? First, do not wait for headlines. The “no news” quarter with a subtly rewritten risk section is the window the market underprices. Second, read insurers differently from utilities. In power, look for procedural nouns and agency‑specific references migrating into 10‑Qs; the linguistic center of gravity moves toward formal process. In insurance, look for modal verbs and qualifiers migrating into capital and tax paragraphs; the center of gravity moves toward balance‑sheet contingency and optionality. Third, connect the filing clock to the event clock. The crown‑jewel cohort’s construction—10‑Qs with rising negative and uncertainty language and a material 8‑K within 90 days—codifies that these are not free‑floating worries. They bracket real events.
The contrast with deep dive A is instructive. In the utilities example, the linguistic signal is an expansion and re‑architecting of risk sections: new subsections, new agency names, more procedural steps. The business meaning is “we are in a process maze.” In insurance, the signal is a tightening and hedging of capital language without necessarily expanding structure: more “may/could,” more explicit reference to buffers and thresholds, and more specificity about where shocks would land. The business meaning is “we are protecting the balance sheet against a policy‑induced distribution of outcomes.” Both changes are reactions to the same macro force—policy uncertainty and enforcement variability—but they propagate through different parts of the filing and carry different implications for timing. Processes resolve on administrative calendars; buffer language resolves on financing and renewal calendars. The trade‑able insight is to recognize the form as well as the magnitude of drift.
Skeptics might ask whether these are simply sector norms rather than signals. Two points. First, the design deliberately focuses on change, not level. A capital‑dense insurer will always talk about RBC; the question is whether the density and hedging rose relative to its prior quarter. That change, not the static category, is where the information resides. Second, the effect’s statistical strength in the crown‑jewel cohort matters. A +6.72% average 60‑day return with a t‑statistic of 11.95 (p<0.001) across 850 observations, with a secondary unit‑beta FF5 control still positive (+1.99%, t=3.17, p<0.01), clears a high bar. It is not a backtest ghost confined to a style box or a single era; it is a language‑first screen that continues to hold after standard controls.
There are also important non‑implications. We are not claiming that every insurer 10‑Q with more modal caution presages a capital action; policy does not always bite. We are not claiming that news sentiment explains the effect; our news program’s standalone alpha was killed and remains conditioning‑only. And we are not claiming that earnings‑call tone carries the same information; that program was trialed and closed null. The point is narrower: when policy drift is quiet and institutional, the earliest reliable trace for portfolio purposes has been in the semantics of the 10‑Q—and pairing that with the 8‑K clock is what turns it into a tradable cohort.
For practitioners, the implication is operational. Build a habit of reading the insurance 10‑Q as a capital document, not just an earnings update. Look for the modal verbs and qualifiers that were not present last quarter; track the emergence of specific risk taxonomies as management breaks out exposure detail it previously summarized; watch for the insertion of phrases that jump from “will” to “expects” to “may,” especially in capital and tax sections. Treat those as the first move. Then anchor a watchlist around the 90‑day event window that follows. If a material 8‑K arrives, the historical return path from the 10‑Q date has not been neutral for this pattern—+6.72% on average at 60 days (t=11.95, p<0.001)—and the effect has survived a standard factor control in our secondary check.
Finally, the conceptual lesson. Language is not a sideshow to numbers; it is how management internalizes policy risk before the ledger moves. In utilities, that internalization reads like process management. In insurers, it reads like balance‑sheet risk management. The investor’s job is to recognize which signature is in front of them and to respect what it implies about the next 90 days. The market will continue to watch hearings and headlines. The companies who live with the policy grind will continue to write 10‑Qs. The spread belongs to those who read them in time.
RESULTS
Executive summary — Quiet Policy Risk Is Migrating Into Routine quarterly reports:
Across the live corpus, return‑relevant disclosure changes are showing up more often inside ordinary quarterly filings rather than being confined to one‑off events. The evidence base below compiles locked results only, with figures drawn from our canonical records. The through‑line is simple: steady‑tone quarterly reports that quietly shift their narrative are where information is concentrating, and the effect is tradable and additive alongside non‑filing signals.
Core quarterly‑report evidence, locked:
Additive, orthogonal signal confirming the migration thesis:
Workforce and sector‑specific corroboration:
Why this matters for Sunday institutional readers:
Risk control and implementation notes (results‑only):
Contextual read of “Quiet Policy Risk” in the numbers:
What this does not say (guardrails and blind spots):
IP tokens: 0
Figures aligned: 100%
As-of: 2026-08-09T06:53:23Z
Allocators don’t need a grand proclamation to act on this; they need a disciplined way to translate a quiet policy signal into portfolio risk that is measurable, diversifying, and falsifiable. The disclosure channel we’ve documented is precisely that. The cohort that matters is simple to define in operational terms and hard to spot in narrative flow: routine 10‑Qs that register a concurrent rise in negative and uncertainty language and that are followed or bracketed by a material event within a 90‑day window. In our locked canon this cohort’s mean 60‑day return is +6.72% with a t‑statistic of 11.95 (p<0.001; n=850), and a secondary specification shows the effect surviving standard factor controls (unit‑beta FF5 residual +1.99%, t=3.17, p<0.01). That is sufficient statistical footing to treat the drift‑plus‑event configuration as a bona fide sleeve rather than an anecdote.
Positioning this sleeve
Treat the drift‑plus‑event cohort as a rules‑based, time‑bounded exposure that sits alongside your core factor and thematic allocations. Because the entry condition is document‑driven and timestamped to public filings, the sleeve is administratively cheap and compliance‑clean. The operating rule is not a directional recommendation on any single issuer; it is a cohort admission rule. When a company’s 10‑Q exhibits a concurrent rise in negative and uncertainty language and a material 8‑K lands within 90 days, that issuer qualifies for the sleeve; when the 60‑day window elapses, it exits. Historical evidence says the expected path of those positions is statistically distinguishable from baseline (t=11.95, p<0.001) and remains so after controlling for market, size, value, profitability, and investment factors in a unit‑beta framework (t=3.17, p<0.01). The practical implication is that you can size the sleeve by risk budget rather than conviction.
Implementation choices cascade from your objective function. If your mandate is absolute‑return, you can run the sleeve long‑only and control gross via a max sleeve weight and a per‑issuer cap. If you’re a market‑neutral allocator, you can implement a factor‑neutralized version with standard FF5 hedges; the point is not to out‑engineer the alpha, but to keep orthogonal risk from muddying attribution. Either way, treat transaction costs conservatively; the sleeve is anchored on SEC timing, not intraday microstructure, and its validation horizon is 60 trading days. Our canon does not rely on newsfeeds or earnings‑call features—the news leg tested as conditioning‑only with no standalone alpha (t=−0.69, p=0.49 under clustering)—so there is no need to pay for speed you don’t exploit.
How to operationalize monitoring
The work is mostly upstream: get your calendar and watchlist right. First, screen fresh 10‑Qs for a rise in both negative and uncertainty language relative to the company’s prior comparable filing. Second, track material 8‑Ks in the 90‑day bracket. You are not forecasting the 8‑K; you are registering that when it brackets the drift, the post‑filing return path has historically diverged from baseline with high significance (t=11.95, p<0.001). Third, maintain a simple ingress/egress ledger for the 60‑day clock. This is not a frenetic strategy; it’s a cadence strategy. Volume is episodic and tied to the filing cycle.
Two corollaries sharpen the allocator playbook. One, don’t condition on headline sentiment. The settled verdict on news is clear: the news module adds conditioning value but no standalone spread (t=−0.69, p=0.49), and attempts to substitute or amplify the filing signal with news sentiment failed cleanly. Two, don’t expect earnings‑call language to substitute for filings. Five pre‑registered variants of a call‑language program cleared none of the gates and the program is closed; the filing channel remains the canonical path. Stated positively: you can run this sleeve on public filings alone, without paying for incremental data that has not demonstrated independent contribution.
Portfolio context and stackability
An allocator’s next question is orthogonality. Our organism’s settled tests show that the filing‑drift sleeve is orthogonal to at least one validated non‑filing signal: legally mandated trading disclosures by members of Congress. When we tested the interaction, the congress‑buy effect persisted both when drift was present and when it was absent (p=0.000 and p=0.0067, respectively), and the contrast between states was statistically indistinguishable (−0.26%, p=0.575). Translation: the two effects do not cannibalize each other and can be stacked. For portfolio construction, that means you can add the drift‑plus‑event sleeve as a separate return stream without assuming it is a proxy for an attention or policy‑flow trade already in your book. The practical implication is better risk‑budget utilization: you are buying independent return drivers, not re‑leveraging one.
Where to watch next
This is a policy transmission story, not a style box. The sectors worth added vigilance are those where licencing, enforcement, and rulemaking cadence shape operations more than incremental demand elasticity. Utilities and independent power producers rewriting risk‑factor narratives around grid policy and permitting, insurers adjusting capital and tax disclosures, logistics incumbents caveating service obligations—these are not meme moments; they are administrative signals. The time to pay attention is the linguistic first move. For a CIO or risk committee, “watch next” breaks into three concrete tasks: (i) sector heat‑maps of drift incidence to see where policy friction is accumulating, (ii) peer‑cohort comparisons within industries to separate idiosyncratic disclosure from sector‑wide boilerplate, and (iii) auditor cadence around these issuers, particularly when filing language hardens. None of those require new data spend; they require discipline in reading what is already public.
Risk management and falsification
Any allocative move is only as strong as its kill switch. This thesis fails if three conditions materialize together. First, if the drift‑plus‑event cohort’s spread collapses toward zero in forward windows so that the 60‑day effect is no longer statistically distinguishable from baseline (e.g., t‑stat falls toward 2.0 with p≥0.05 across fresh cohorts), we retire the sleeve. Second, if factor adjustments absorb the effect—i.e., the unit‑beta FF5 residual shrinks to insignificance (t≈0, p≥0.05)—then what looked like policy‑channel alpha was just factor exposure in disguise and should be de‑risked. Third, if an across‑the‑board change in disclosure norms inflates uncertainty and negative language without corresponding event follow‑through, the base rate of drift flags will rise while the ex‑post event frequency falls; the cohort will bloat and its economics will vanish. We have a settled verdict that the 2026 “disclosure elevation” hypothesis was a null/artifact, which is a guardrail against this concern; but the correct stance is to keep auditing base rates and maintain a low tolerance for degradation.
Two additional cautions deserve explicit policy. First, avoid smallest‑liquidity tails when you implement. Our validation work emphasizes reliability above cleverness; the effect does not require reaching into the thinnest names to show up. Second, do not improvise new conditioning layers ad hoc. The temptation, when a signal works, is to make it “better” with unrelated features. Our news and earnings‑call findings make clear that adding unvalidated layers dilutes statistical power without adding return.
Governance for adoption
Institutional adoption is smoother when the evidence chain is simple and the operational touchpoints are few. That is the case here. The evidence chain has three links with statistics attached: (1) document‑level drift in routine 10‑Qs concentrated in negative and uncertainty language; (2) a material 8‑K bracketing within 90 days; (3) a 60‑day forward return profile that is statistically different from baseline (t=11.95, p<0.001), including after FF5 controls in unit‑beta space (t=3.17, p<0.01). The governance corollary is straightforward: pre‑register your admission criteria, your holding window, and your risk caps; track adherence; and report attribution with those same three links. If you run multi‑sleeve portfolios, record the orthogonality tests (p=0.575 for the congress‑x‑drift contrast) and keep them current.
What success looks like in practice
Success is not “calling” specific enforcement or legislative events. Success is installing a low‑touch, statistically defendable sleeve that clips a repeatable, policy‑linked drift in public filings that other allocators do not watch until it is too late. If you can report, quarter over quarter, that (i) the number of admitted issuers stayed within risk budget, (ii) the realized 60‑day cohort return stayed within the validated confidence bands implied by t=11.95 (p<0.001), and (iii) factor‑neutral attribution remains positive with residual significance mirroring t=3.17 (p<0.01), then the sleeve earns its keep. If those conditions fail, the kill switch triggers and capital redeploys. That is how intelligence becomes allocation without theatrical forecasts.
The bottom line is that policy risk in 2026 is migrating through the disclosure channel before it hits guidance or headlines. You do not need a new class of data to exploit it; you need a ruleset disciplined enough to listen to the first move—language—and cheap enough to ignore the noise—news—until the event lands. The work is mostly patience and record‑keeping, which is not glamorous but is scalable. In the taxonomy of allocator edges, that qualifies as a quiet one worth taking, because the statistics meet the bar today (t=11.95, p<0.001; FF5 residual t=3.17, p<0.01) and the governance is as simple as an admission rule and a clock.
This appendix sets out the validated signal set that bears directly on the theme of policy risk migrating into routine 10‑Qs. It is a record of what our data supports, and where the limits are. All figures cited are from locked canon, and where we lack a validated number we make no claim. The common thread is simple: the market’s attention to headlines is episodic, but the disclosures that quietly foreshadow those episodes leave measurable traces in filings.
The crown‑jewel cohort
Our primary evidence is the cohort we refer to as the crown jewel. It is narrowly defined: quarterly filings in which the language shows a concurrent rise in negative and uncertainty terminology, bracketed by the presence of a material event in proximity. We do not need to infer why management wrote differently; we only need to observe that they did, and that a material 8‑K lands within a 90‑day window around that 10‑Q. The return leg after the filing is not neutral in this cohort. The locked figure is a mean 60‑day return of +6.72% with a t‑statistic of 11.95 on n=850 observations. A secondary specification that controls for standard Fama‑French factors yields a positive residual effect as well (+1.99%, t=3.17). That is the core of the theme: a measured linguistic shift in routine disclosure, co‑occurring with a material event, and an ensuing return path that diverges from the baseline.
The crown‑jewel result makes three things clear. First, the locus is in the filing language, not in guidance or contemporaneous newsflow. The cohort is constructed from the 10‑Q text and the simple fact that a material event lands within the defined window. Second, the mechanism is about change from baseline, not absolute levels. Firms with historically cautious language are not mechanically pulled into the cohort; it is the shift in uncertainty and negative density that matters. Third, the signal is not a clairvoyant alarm for policy or legal events; it does not “predict” the 8‑K in a causal sense. It tells us that when the text and the event coincide in this way, the subsequent 60 trading days are statistically different from what the market treats as ordinary at the time of filing.
Equally important is what the crown‑jewel cohort does not claim. It is not a broad regime diagnosis of “elevated disclosure” across the market; a registered attempt to generalize a 2026 disclosure‑elevation trend tested as a null artifact and was closed as such. It is not a news overlay; a stand‑alone news stream and overlays on filing cohorts tested cleanly null in our adjudicated work. It is also not an earnings‑call substitute. A pre‑registered program that attempted to transplant the filing‑language machinery onto earnings calls was closed after five variants failed to produce a validated signal. The crown‑jewel result is specific: filings, not calls; language change, not boilerplate; co‑occurrence with a material 8‑K, not an abstract anticipation of one.
Congressional buying
The second line of evidence relevant to the policy‑risk thesis comes from the congressional trading record. Two variants of the congress‑buy signal cleared our validation battery and are locked. The trade‑date variant, anchored to the date the member transacts, delivers +1.96% with p=0.007 (n=4,903). The disclosure‑date variant, which is the tradable version because it keys to public information, delivers +2.66% with a confidence interval of [+1.1%, +4.3%] (n=4,101). A member‑weighted version strengthens the signal to +3.11% with p<0.001. These are not filings; they are public disclosures of personal trades. They matter here because they speak to the institutional channel through which policy risk can transmit into prices.
The interaction with filing‑language signals is resolved, and the result is practical: independence, not conditioning. A pre‑registered contrast test found that the congress‑buy effect does not materially differ by the state of disclosure‑language drift; the contrast was statistically null, and the effect persists in both states. The implication is straightforward. Where the crown‑jewel cohort tells us that language drift in a 10‑Q plus a nearby material event shapes the subsequent return path, the congress signal tells us that elected officials’ trading disclosures carry their own tradable edge. The two signals are uncorrelated and stackable; they do not need each other to work, and neither is a filter for the other. That orthogonality matters for portfolio construction and for interpreting what the filings are telling us about slow policy movement. It supports the central theme: policy risk is not flashing in headlines alone; it is showing up in how insiders act and in how management writes.
Workforce–WARN convergence
A third validated signal that bears on the theme links filings to the labor channel through administrative disclosures. The workforce–WARN convergence signal identifies firms where workforce‑reduction alerts in the administrative record coincide with shifts in disclosure language. It is validated in our ledger and therefore may be cited as operational. We do not publish statistics beyond the status because the canon does not fix a single external metric for this signal. What matters for the current theme is the construct: when management begins to write with greater uncertainty and negativity in routine risk‑factor updates, and the workforce record points to structural adjustments, the two move together often enough to be useful. The slow policy grind described in the theme – rulemaking backlog, enforcement variability, licensing and tax ambiguity – frequently registers first as administrative adaptation (headcount, hours, locations) and linguistic caution. The convergence signal captures that pairing.
As with the other validated pieces, it is critical to stipulate the boundaries. Workforce–WARN convergence is not a return‑targeted claim in this appendix; it is a risk identification mechanism. It does not say that any given WARN cluster implies a policy‑driven outcome in the stock. It says that when the workforce administrative trail and the 10‑Q language both lean in the same direction, the firm is moving through an adjustment phase that deserves attention. In the context of utilities revising licensing risk language, insurers adjusting tax and capital sections, or logistics incumbents caveating service obligations, that convergence is often the first visible evidence of a slow policy‑to‑operations transmission.
What the data do not support
Rigor requires being explicit about the dead ends and the boundaries so that the claims above are not read beyond their scope. Three settled verdicts are directly relevant to this theme. First, news does not stand alone as an additive alpha source in our framework. A condition‑only role for news remains – one can use it to define windows – but a stand‑alone or overlay news alpha in the presence of filing‑language cohorts tested null across replications. Second, the attempt to port the filing‑language approach to earnings‑call transcripts has been closed; no validated call‑side signal emerged in the pre‑registered program. Third, the temptation to call a broad market‑wide elevation in disclosure tone in 2026 was adjudicated as an artifact and killed. These verdicts matter because they keep the focus on what the filings actually do: they are the substrate where management encodes risk and uncertainty language, and where we see the policy creep first.
There are also open questions on which we make no claim here. We do not assert any anomaly in onset rates for the current quarter; that work is explicitly open and therefore out of scope for external statements. We do not present a stand‑alone offering‑language signal; a historical attempt to do so was killed as a null. And we do not present a boilerplate‑driven “verbosity” shortcut; those routes have also been closed as artifacts when subjected to factor‑controlled validation. The lesson is consistent: where adjudication is incomplete or null, it stays out of this appendix.
Implications for the theme
Taken together, the validated signals draw a coherent picture of how quiet policy risk migrates into prices. The crown‑jewel cohort demonstrates that when a 10‑Q’s language turns more uncertain and negative and a material event is in the orbit, the next two months look different in the tape. The congressional signal demonstrates that there is a second, institutional channel through which policy information becomes tradable, and that it is independent of how management writes its filings. The workforce–WARN convergence shows that, in many cases, internal operational adjustments and administrative notices begin to align with the change in language before markets have repriced the full consequence.
That picture has practical consequences for investors and risk managers. First, the actionable surface is in the filings. You do not need to chase earnings calls or pay for premium newsfeeds to see the change the moment it registers. The numbers above are validated under standard factor controls and are constructed from public data. Second, the signals are not substitutes for one another. The congress‑buy edge can sit alongside filing‑language signals because the two are uncorrelated. That is a portfolio design point, but it is also a substantive one: multiple channels are carrying policy information into markets, and they do not cancel each other out. Third, the operational path matters. A firm that has begun to caveat service obligations, to modify licensing risk language, or to add modal cautions in its risk factors while its administrative workforce record tilts in the same direction is in a policy‑induced adjustment period. Whether or not the 8‑K has printed yet, the filings are telling you that the story has already started.
Finally, we reiterate what this appendix does not do. It does not generalize beyond the validated constructs. It does not claim a new regime statistic for this quarter. It does not assign causation to the language changes. It stays within locked canon: the crown‑jewel cohort’s +6.72% mean 60‑day return (t=11.95, n=850) with a positive factor‑controlled residual; the congress‑buy signal’s traded and tradable edges (+1.96% trade‑date; +2.66% disclosure‑date with a defined confidence interval; a member‑weighted +3.11% variant); and the validated status of workforce–WARN convergence as a risk‑identification mechanism. Those points are sufficient to support the theme’s conclusion: the policy transmission in 2026 is less about spectacle and more about the slow, linguistic and administrative grind that ordinary 10‑Qs faithfully record.