Mandatory Execution Confirmation (Form 144-A)

• Mandatory Execution Confirmation (Form 144-A) is a proposed SEC disclosure that links initial intent notices with actual trade executions.
• It addresses the reporting inversion by clarifying whether insider sale plans are fully executed, partially executed, or aborted.
• Empirical studies and machine learning analyses support that making execution outcomes public can reduce market opacity and improve compliance.

Searching arXiv for the cited papers and closely related Form 144/Form 4 disclosure work to ground the article. Tool call: arxiv_search({"query":"id:(Huang et al., 27 Jul 2025) OR id:(Neupane, 19 Feb 2026) Form 144 Form 4 insider reporting inversion Mandatory Execution Confirmation", "max_results": 10, "sort_by": "relevance"}) Mandatory Execution Confirmation (Form 144-A) is a proposed supplemental SEC disclosure designed to close the informational gap between a filed Form 144 notice of proposed sale and the eventual execution or non-execution of that proposed sale. In the literature, the proposal emerges from two adjacent findings: first, that the post-Sarbanes–Oxley disclosure regime has produced a “reporting inversion” in which Form 4 often precedes or effectively overtakes Form 144 in informational relevance; second, that modern compliance modeling can translate deadline-based disclosure rules into large-scale, transaction-level surveillance tasks with high empirical accuracy when the underlying labels are observable (Neupane, 19 Feb 2026, Huang et al., 27 Jul 2025). Form 144-A is therefore best understood not as a standalone filing innovation, but as a mechanism for converting an unresolved intent signal into a verifiable outcome record.

1. Regulatory position within the Form 144 and Form 4 disclosure system

Form 144 is the “notice of proposed sale of securities” under Rule 144. It is filed by insiders and other affiliates before or at the time of placing an order to sell restricted or control securities, and it opens a 90-day execution window during which the filer may sell up to the number of shares stated in the filing. Form 4, by contrast, is the transactional insider trading report under Section 16(a); it is ex post and must be filed within two business days after the trade date of an insider transaction in the issuer’s equity securities (Neupane, 19 Feb 2026).

The distinction is foundational. Form 144 discloses intent and capacity to sell, but not a binding commitment to execute. Form 4 discloses an actual execution, including transaction date, number of shares, price, and ownership details. The proposed Form 144-A would sit after both the initial notice and the execution window, disclosing how the original intent resolved: fully executed, partially executed, or not executed.

Instrument	Timing	Function
Form 144	Before or at the time of placing an order to sell; opens a 90-day window	Notice of proposed sale of restricted or control securities
Form 4	Within two business days after the trade date	Report of actual insider transaction
Form 144-A	Within a short specified period after the 90-day window expires or the proposed volume is fully executed	Mandatory execution confirmation or abort disclosure

This arrangement matters because the current regime is asymmetric by design. Form 144 states that a sale may occur, but does not require any ex post confirmation that it did occur. Form 144-A is proposed precisely to supply that missing state variable.

2. Reporting inversion and predictive decoupling

The central institutional problem is the post-SOX “reporting inversion.” Before Sarbanes–Oxley, Form 144 preceded Form 4 in 93.7% of pre-SOX cases. Post-SOX, Form 144 preceded Form 4 in only 17.5% of cases. The practical result is that the market often learns first that an insider has already sold, through Form 4, and only later sees or reconstructs that the sale had been pre-notified under Rule 144 (Neupane, 19 Feb 2026).

The paper on insider intent characterizes this as “predictive decoupling” between the intent-to-sell signal in Form 144 and the realized execution path. A Form 144 opens a 90-day interval in which the insider may execute up to the proposed volume, but some filings are aborted and some are executed, with no mandatory ex post filing stating which path occurred. That missing confirmation is not a clerical detail; it is the source of a structural informational friction.

The formalization is explicit. For a filing $i$, the paper defines non-execution as:

$$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$

Under the present regime, this binary outcome is not itself an official public disclosure. Market participants must infer it indirectly by matching Form 4 executions to Form 144 notices over the 90-day window. The paper argues that this induces “non-symmetric pricing” and violates the symmetric-information condition invoked through the First Fundamental Theorem of Welfare Economics. Prices during the 144 window embed a probabilistic discount for possible insider selling, but the market cannot reliably distinguish opportunistic informed sales, routine liquidity sales, and aborted sales driven by private information.

A common misconception is that Form 144 already provides sufficient transparency because it is public and ex ante. The evidence presented in the paper rejects that view: Form 144 discloses only a proposal, not a realized state. The unresolved state is the precise source of the “predictive blind spot” that Form 144-A is intended to remove.

3. Empirical evidence for the informational gap

The empirical case for Form 144-A rests on event-study, portfolio, and machine-learning evidence. The paper defines the gap between filing and execution as

$$G_{\text{days}, i} = \text{Date}_{\text{Execution}, i} - \text{Date}_{\text{Filing}, i},$$

and computes cumulative market excess return over the live interval between Form 144 filing and execution as

$$CMER_i = \left[ \prod_{t=F+1}^{E} (1 + R_{m,t}) \right] - 1.$$

Across 2.6 million trades, mean $CMER$ during the filing gap is 1.6484%, with $t = 553.93$ and $p < 0.001$. The paper further reports that insiders scale sales with market conditions: in low-CMER environments, mean signal magnitude is approximately 1.63%, while in high-CMER environments it rises to approximately 51.13%. This is interpreted as insiders scaling up large, high-information sales when broader market appreciation can mask the firm-specific negative signal (Neupane, 19 Feb 2026).

The event-study results sharpen the point. Using a Carhart four-factor framework, the paper reports $CAR_{144}(0,+2) \approx 0.16\%$, $CAR_{4}(0,+2) \approx 1.45\%$, and a paired difference of $-1.2943\%$ with $$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$0 and $$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$1. The paper treats this roughly 1.3% differential as an execution information premium: the market reacts more strongly around execution than around intent disclosure, implying that the informational content of the proposed sale is not fully impounded at the Form 144 stage.

The machine-learning audit is equally important. Even with SMOTE and heavy cost weighting, recall for aborted trades never exceeds approximately 50%, and PR-AUC for the best performers tops out around 0.42. For the benchmark XGBoost classifier, 46.77% of aborted intents are misclassified as executions, yielding a “Strategic Opacity” floor of approximately 52.4%. The paper describes this plateau as a “structural information ceiling,” not merely an algorithmic shortcoming, because advanced tabular and neural models converge to similar limits.

Cross-sectional evidence extends the argument. Size-sorted portfolios show a “large-cap significance paradox”: small-cap portfolios yield higher absolute abnormal returns, 32.21 basis points in one summary and 30.96 basis points per month in the median split, but statistical significance is concentrated in large-cap firms, where alpha is 14.49 basis points per month with $$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$2. Prior idiosyncratic volatility is a significant amplifier, with coefficient $$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$3 and $$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$4. Causal estimators further identify illiquidity jumps of up to 2.63 times for large intended trades. Taken together, these results imply that the missing execution-status disclosure is not a minor reporting nuisance; it is associated with persistent opacity, delayed repricing, and measurable liquidity effects (Neupane, 19 Feb 2026).

4. Structure and purpose of the proposed Form 144-A

Form 144-A is proposed as a mandatory execution confirmation filed by the same insiders or affiliates who filed the original Form 144. The filing would occur within a short specified period after either the expiration of the 90-day window or the full execution of the proposed volume, whichever comes first. Its core content would be a direct mapping from the original Form 144 to subsequent outcomes: whether the proposed sale was fully executed, partially executed, or not executed, together with executed volume versus proposed volume and either weighted average execution price information or references to the relevant Form 4 entries (Neupane, 19 Feb 2026).

The proposal is framed as a transition from unilateral disclosure to bilateral accountability. Under the present system, insiders declare intent but are not required to confirm whether they followed through. Form 144-A would require them to disclose the resolution state. That change would make every Form 144 part of a fuller chain:

Form 144 $$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$5 Form 144-A $$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$6 possibly linked Form 4 entries.

The policy rationale is narrow and targeted. Form 144-A does not replace existing anti-fraud or insider trading provisions, and it does not eliminate all insider informational advantages. Its specific function is to make the latent binary state—execute versus abort—observable. This suggests that the proposal is less a substantive change in who may trade than a change in the observability of trade-plan resolution.

The enforcement implications follow directly. Failure to file Form 144-A, or materially inaccurate reporting of execution status, would become a compliance violation. Regulators could then identify insiders who repeatedly file large Form 144 notices without executing, or who sequence notices and executions in suspicious ways around material events. The paper presents this as a way to reduce “phantom risk” premia and restore informational symmetry in the capital market.

5. Computational surveillance and the IFD benchmark as a transfer model

A second line of work shows how deadline-based securities disclosures can be operationalized as transaction-level compliance detection. The IFD benchmark, “Insider Filing Delay,” is built from WRDS Form 4 transactions and SEC data and contains 4,051,143 trade-level records from 2002 to 2025, covering more than 7,633 U.S. public companies and 15,573 individuals. It includes 52 attributes grouped into identification/date fields, company and security information, form and transaction details, data-cleansing indicators, timing and delay information, and additional fields such as a Rule 10b5-1 plan flag, market value of transaction, proposed number of shares for Form 144, and a binary violation label (Huang et al., 27 Jul 2025).

In IFD, Form 4 timeliness is translated into a hard binary label under the SEC’s two-business-day rule:

$$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$7

The dataset further defines historical violation frequency by insider and by firm:

$$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$8

and

$$\text{Non-Execution}_i = \begin{cases} 1 & \text{if } \sum V_{executed,i} < V_{proposed,i} - \epsilon \ 0 & \text{if } \sum V_{executed,i} \geq V_{proposed,i} - \epsilon \end{cases}$$9

This framework supports a supervised binary classification formulation over structured transaction data, combining insider identity and role codes, trade timing, filing timing, security identifiers, trade size and value, ownership flags, and governance and financial indicators. The benchmark reports that roughly 17.4% of trades are classified as violations; about 77% of violations are characterized as “oversight” and about 23% as “intentional.” Beneficial owners exhibit a notably higher role-based violation rate, 22.24%, than CEOs or the corporate suite. Supplementary analyses also identify 56,000 trades in “stealth sequences,” 697 round-trip trades, and approximately 43,000 trades within 60 days before earnings but disclosed only after the announcement (Huang et al., 27 Jul 2025).

The proposed model, MaBoost, combines a Mamba-based state-space sequence encoder with XGBoost. On Form 4 delayed disclosure detection, it reaches an F1-score of up to 99.47% under the “Constraint Condition,” with reported precision of 99.09 and recall of 99.85. Feature ablations show that insider history, spatiotemporal timing, governance, and trade characteristics are especially consequential, while the authors emphasize the interpretability advantages of a tree-based back end.

This suggests a direct implementation path for Form 144-A surveillance. If a jurisdictional rule defines a confirmation deadline relative to execution or window expiration, then an analogous label can be written as

$$G_{\text{days}, i} = \text{Date}_{\text{Execution}, i} - \text{Date}_{\text{Filing}, i},$$0

with

$$G_{\text{days}, i} = \text{Date}_{\text{Execution}, i} - \text{Date}_{\text{Filing}, i},$$1

A plausible implication is that Form 144-A would not only improve market transparency but also enable benchmark-quality compliance datasets of the kind already demonstrated for Form 4, because the currently latent execution-status label would become explicit.

6. Significance, limits, and unresolved questions

The significance of Form 144-A lies in its attempt to resolve a specific market failure documented in the literature: the inability of public data to distinguish executed from aborted intent signals at the event level. By making that resolution observable, the proposal is expected to reduce the 1.3% execution information premium, attenuate post-expiration abnormal returns associated with aborted intents, and narrow the informational asymmetry that underlies the reported 52.4% opacity floor (Neupane, 19 Feb 2026).

At the same time, the proposal has clear limits. It does not remove all private information held by insiders. It does not replace anti-fraud rules, insider trading prohibitions, or the broader compliance infrastructure around Rule 144, Form 4, and Rule 10b5-1 plans. The literature also identifies practical caveats relevant to any operational deployment: label noise from misrecorded dates, reporting biases across time, regime shifts in disclosure rules, and the risk of over-reliance on highly accurate but dataset-specific models. The IFD results themselves caution that very high F1 scores are specific to the benchmarked Form 4 task and that other domains, including Rule 144 or 144A settings, may be sparser and therefore harder (Huang et al., 27 Jul 2025).

Several open questions follow directly from the existing research. One is behavioral adaptation: insiders may file fewer but more committed Form 144 notices if Form 144-A is adopted. Another is institutional design: the literature raises, but does not fully settle, the optimal penalty structure for missing or inaccurate confirmation filings. A third concerns integration with preplanned trading programs and other compliance workflows. A fourth is comparative: the evidence invites cross-jurisdictional work on whether other markets exhibit analogous “information ceilings” when intent disclosures are not paired with outcome confirmations.

In integrated terms, Mandatory Execution Confirmation (Form 144-A) is a proposal to transform Form 144 from a one-sided announcement of possible action into a completed disclosure lifecycle. The empirical literature motivates it through reporting inversion, predictive decoupling, abnormal return patterns, and a machine-learning ceiling on inferring the abort-versus-execute state from public observables. The compliance literature shows that once such states are formally labeled, they can support large-scale, interpretable surveillance systems. The proposal’s core contribution is therefore epistemic as much as regulatory: it seeks to replace an unresolved latent state with a verifiable public record (Neupane, 19 Feb 2026, Huang et al., 27 Jul 2025).