Comparing Insider Transaction Disclosure Rules: US vs EU

Note on figures cited below. Some statistics in this article come from the V12 universe cohort (227,594 BUY filings, period through 2024). Since 2025-01-01 the live out-of-sample BUY universe (28 markets) shows a 52.96 percent win rate and a +8.59 percent mean 90-day return on 20,688 filings. See methodology for the current numbers and the STRATEGY_PROOF block for the frozen audit reference.

A filing clock is only as useful as the regulator willing to read it.

The phrase "3-business-day rule" sounds crisp, almost comforting. Then you look at the filings.

The live extract you provided is enough to make one point with unusual clarity. Cross-regulator comparisons are not hard because the law is vague. They are hard because the data are dirty, the clocks differ, and enforcement is uneven. A market-level average that says publication happened 181.5 days before the transaction is not a subtle statistical issue. It is a sign that some combination of venue mapping, date semantics, or ingestion logic is wrong.

That is not an embarrassment. It is the story.

What the rule says, and what it does not say

The legal promise of T+3

Across major markets, insider transaction disclosure rules are designed to reduce the period in which management knows something the market does not. The broad idea is familiar. The details are not.

In the European Union, Article 19 of the Market Abuse Regulation, MAR, requires persons discharging managerial responsibilities, PDMRs, and closely associated persons to notify the issuer and the competent authority of transactions conducted on their own account. The issuer or market participant must then ensure public disclosure, and the operative timeline is generally framed around three business days after the transaction date. ESMA has supplemented this with Q&A and technical standards on notification templates and thresholds.

In the United States, Section 16 insiders report changes in beneficial ownership on Form 4, generally by the end of the second business day following the transaction date. The SEC's deadline is, on paper, tighter than MAR's common T+3 framing.

If this were only a matter of statutory text, the ranking would be easy. SEC should look faster than European supervisors. Yet real-world publication lag is not just a function of deadline length. It is a function of filing rails, machine readability, publication architecture, and whether late filers face consequences beyond a raised eyebrow.

Enforcement asymmetry is the real variable

The asymmetry comes in several forms.

First, centralisation. The SEC's EDGAR system is a single, highly standardised publication mechanism. Once accepted, the filing is public almost immediately. In Europe, MAR is harmonised at the regulation level, but supervision and publication still run through national competent authorities and issuer channels. Harmonised law, fragmented plumbing.

Second, visibility of lateness. In the US, delinquent Section 16 reporting has long been a visible governance issue, reinforced by annual proxy disclosure under Item 405 of Regulation S-K. Public shaming is not a complete enforcement regime, but it is a regime. In Europe, the sanctioning framework exists under MAR, but the practical intensity and publicity of enforcement vary by country.

Third, data structure. SEC Form 4 is machine-readable XML and timestamped at filing. European disclosures often arrive in regulator bulletins, issuer announcements, PDFs, or feeds with inconsistent field definitions. That makes "event to publication" harder to measure consistently, even before anyone misses a deadline.

The result is that comparing median lag by regulator is less like comparing speed limits and more like comparing roads, police, and odometers at once.

Your dataset already shows why naive comparisons fail

Negative lags are not a quirky edge case

The query supplied calculates average days between pubDate and transactionDate by market. The first rows are spectacularly negative: HAYN at -181.5, LPK at -24.6, GUG at -10.5, LEA at -2.4, followed by a long tail of -0.5 values.

A publication date that precedes the transaction date can happen in exactly three broad ways:

1. The wrong event date is being used.
   For example, a filing may report a transaction execution date, a deemed execution date, a grant date, or a settlement date. If your parser grabs the wrong one, chronology breaks.

2. The wrong publication timestamp is being used.
   Some systems store local publication date, UTC timestamp, or feed ingestion time. A same-day filing can turn into -0.5 days if one field is date-only and the other is midnight-normalised in another timezone.

3. The grouping key is not a regulator.
   Your market field appears to contain a mixture of actual markets, venue labels, and what look suspiciously like tickers or instrument codes. HAYN and LEA are also US-listed tickers. That is not a jurisdictional regulator comparison. It is a taxonomy accident wearing a regulator's hat.

The -0.5 cluster is especially telling. It smells like date truncation or timezone conversion, not mass clairvoyance by insiders.

Average is the wrong first statistic

The article angle asks for median lag by regulator. Sensible. The query provided uses average lag by market. That is a useful smoke alarm, but a poor scoreboard.

Why median? Because filing lag distributions are ugly. They have spikes at zero, one, two, and three business days, plus a long right tail of late filings, corrections, and data errors. A handful of absurd negatives or very late outliers can move the mean more than the median. If the purpose is to compare practical timeliness across regulators, median is the robust baseline, and percentile bands should sit beside it.

A better first pass would be:

• map each filing to a regulator or jurisdiction rather than market,
• exclude or flag records where pubDate < transactionDate,
• compute median calendar lag and, if possible, median business-day lag,
• show p25, p75, and late-filing share beyond the legal deadline.

That would turn a curiosity cabinet into an analysis.

SEC versus MAR, the text is simple, the machinery is not

The SEC has a shorter deadline and better rails

The SEC's Form 4 regime is the benchmark for operational clarity. Insiders subject to Section 16 generally file by the end of the second business day following the transaction. EDGAR provides a central submission and dissemination system, with standard fields, acceptance timestamps, and broad data availability. Researchers can debate economic interpretation. They do not usually debate whether the filing existed.

This matters because enforcement is partly infrastructural. A regulator with a central electronic system does not need to "discover" many late filings. The system itself records the filing time against the transaction date. That creates a natural audit trail and lowers the cost of supervision.

It also helps market participants. Data vendors can parse Form 4 at scale. Investors can monitor timeliness with relatively little manual cleaning. The compliance burden is not trivial, but the information path is legible.

Europe has harmonised law and decentralised practice

MAR was meant to reduce fragmentation, and in legal terms it did. Article 19 applies across the EU, and ESMA has worked to standardise templates and interpretation. But supervision remains national. Publication channels differ. Issuers often play a larger role in dissemination. National competent authorities vary in their technical interfaces, archive quality, and sanctioning style.

France's AMF is a good example of a serious regulator operating inside this decentralised architecture. The rule is clear enough. The publication path is less uniform than EDGAR. Depending on the issuer and the feed captured by your database, pubDate may reflect regulator publication, issuer release, vendor ingestion, or a normalised timestamp generated downstream. Each option is defensible. None are interchangeable.

This is where enforcement asymmetry becomes practical rather than theoretical. Two regulators can both endorse T+3. One may have a highly visible, machine-auditable pipeline and routine sanction notices. Another may rely more heavily on issuer compliance and ex post supervisory action. The law on the page is the same. The observed lag in your dataset may not be.

Sanctions exist, but comparability does not come for free

Both the SEC and European authorities can sanction non-compliance. The issue is not whether powers exist. It is how often they are used, how visible they are, and whether they create a credible expectation of timely filing.

The SEC's ecosystem benefits from repeated public attention to Section 16 compliance, including proxy disclosure of delinquencies. Under MAR, administrative sanctions are available and public sanction notices do appear, but the intensity and style of use vary across member states. ESMA publishes annual reports on administrative and criminal sanctions under MAR, which are useful for context, though they are not a direct measure of insider filing timeliness.

That distinction matters. A regulator can produce many sanction notices because it is active, or because its market is disorderly, or both. Sanction counts are not a clean proxy for filing speed. They are, however, part of the enforcement backdrop that shapes incentives.

How to measure enforcement asymmetry properly

Start with a regulator map, not a market field

The central analytical fix is boring and therefore important.

You need a regulator dimension table. Each filing should map to:

• jurisdiction,
• regulator or competent authority,
• legal regime, for example SEC Form 4 or MAR Article 19,
• publication channel type, such as regulator bulletin, issuer RNS-like release, EDGAR, exchange feed, vendor-derived.

Only then can you ask the intended question: does the median event-to-publication lag differ by regulator?

Using Company.market is understandable if that is what the schema offers. It is not good enough for this article. The current result set is clearly contaminated by non-regulatory identifiers.

Use business-day lag, but keep calendar lag visible

The phrase "3-business-day rule" invites business-day measurement. Do it. But do not throw away calendar lag.

A Friday trade published on Tuesday can be one business day or four calendar days depending on holidays and jurisdiction. Both are informative. Business-day lag tests compliance against the rule. Calendar lag measures how long the market actually waited.

For a regulation article, the cleanest presentation is:

• median business-day lag by regulator,
• median calendar-day lag by regulator,
• share of filings published after the legal deadline,
• share of impossible negatives,
• sample size after cleaning.

That last line is not decorative. It tells the reader whether the ranking rests on 30 filings or 30,000.

Separate same-day publication from same-day transaction

There is another subtlety. Some disclosures are published the same day as the transaction, especially in highly automated systems or where the trade occurred early enough in the day. Others may appear "same day" because the date field drops the time component. If your data stores only dates, a filing made at 23:55 after a 09:30 trade is indistinguishable from one made before the market opened.

That is why a dense cluster at -0.5 or 0.0 days should trigger a timestamp audit. If time-of-day is unavailable, report the limitation plainly. Better a modest caveat than a heroic fiction.

What the literature says about insider disclosures and timeliness

Markets react quickly when the data are reliable

Academic work on insider trading disclosures has repeatedly found that insider purchases, in particular, contain information and can move prices around disclosure. The exact magnitude varies by market, period, and methodology. The broad point is stable. Timely disclosure matters because the information has value.

In the US, the Form 4 regime has supported a large empirical literature precisely because the data are centralised and timestamped. In Europe, MAR and predecessor regimes have also generated evidence of informational content, but cross-country comparability is often weaker because disclosures are more heterogeneous.

This is where enforcement asymmetry bites investors directly. If one regulator's pipeline produces cleaner and faster publication, the market can incorporate insider information more uniformly. If another jurisdiction's disclosures are slower, patchier, or harder to parse, the informational benefit of the rule is diluted even if the statute looks equally strict.

Enforcement changes behaviour at the margin

There is a second literature, more governance-oriented, showing that disclosure and enforcement shape managerial behaviour. The effect is not magical. Insiders still trade, and some file late. But visible, standardised reporting obligations reduce ambiguity and raise the expected cost of non-compliance.

Again, the practical lesson is not that one should count sanctions and declare victory. It is that a functioning disclosure regime is a system. Deadline, forms, publication rails, auditability, and enforcement all interact.

A T+2 rule with weak infrastructure can underperform a T+3 rule with excellent infrastructure. Equally, a T+3 rule with fragmented publication and uneven sanctions can underperform both. The market does not trade on legal prose. It trades on when information becomes visible.

A workable blueprint for your 162k-filings study

The minimum cleaning protocol

If this article is to compare median lag by regulator using your 162k filings, here is the minimum viable protocol.

1. Create a regulator mapping table.
   Assign each issuer or filing source to SEC, AMF, FCA, BaFin, CONSOB, CNMV, and so on, based on domicile, listing venue, and source feed.

2. Validate date fields.
   Confirm whether transactionDate is execution date, notification date, settlement date, or another event. Confirm whether pubDate is regulator publication timestamp, issuer announcement date, or vendor ingestion date.

3. Flag impossible chronology.
   Keep negative lags in a diagnostics table, but exclude them from headline medians unless you can repair them.

4. Compute both calendar and business-day lag.
   Use jurisdiction-specific holiday calendars where feasible. If not, state that business-day lag is approximated.

5. Report coverage and exclusions.
   Show how many records survive cleaning by regulator. Readers should know whether the final comparison rests on 90 percent of the data or 40 percent.

6. Add lateness rates.
   Median lag is useful. Share filed after the legal deadline is more directly tied to enforcement.

That protocol would let you write a regulation article rather than a forensic note from the database engine.

The chart that would actually answer the question

If you rerun the analysis correctly, the ideal visual is simple:

• one row per regulator,
• median business-day lag,
• p75 lag,
• late-filing share,
• negative-lag share before cleaning.

That final column is not a nuisance. It shows where the data pipeline itself may be weakest. In a dry way, it is also a nice reminder that some regulators are not slower than others, they are merely observed through foggier windows.

What we can responsibly say now

With the current extract, three claims are defensible.

First, the dataset is large enough to support a meaningful regulator comparison once mapped correctly. 162,000 filings is not a toy sample.

Second, the raw result demonstrates substantial data-quality issues that would bias any naive ranking. Negative average lags as low as -181.5 days are disqualifying for headline comparison.

Third, enforcement asymmetry remains the right explanatory frame. Even after cleaning, one should expect practical lag differences across regulators because filing infrastructure, dissemination channels, and sanction visibility differ.

What we cannot responsibly say, yet, is that regulator A has a lower median lag than regulator B in your data. The query does not identify regulators, and the chronology is not clean.

Why this matters beyond compliance trivia

Timeliness changes the value of the signal

Insider transaction data are not merely governance ornaments. They are an information signal. A purchase disclosed promptly can inform price discovery, analyst interpretation, and shareholder oversight. The same purchase disclosed late is still interesting, but less useful. At some point it becomes historical scenery.

That makes enforcement asymmetry economically relevant. If one market's insiders are effectively visible within one or two business days and another market's disclosures drift, the informational playing field is uneven. Researchers will detect different abnormal-return patterns. Investors will build different monitoring habits. Issuers will face different reputational pressure.

The rule is only as credible as the publication path

This is the unglamorous core of the article. Disclosure rules are often discussed as if they are legal abstractions. They are not. They are operational systems. The market experiences them as timestamps.

When a regulator combines a clear deadline, standard forms, central electronic filing, immediate public dissemination, and visible consequences for lateness, the rule becomes credible. When any of those pieces are weak, the deadline starts to look aspirational.

Your extract, accidental though it may be, captures this perfectly. A negative average lag is what happens when the publication path is not yet measured with enough discipline. Before one can judge regulators, one has to judge clocks.

The concrete next step is obvious: rebuild the comparison at the regulator level, not the market field, and publish median business-day lag, late-filing share, and negative-lag diagnostics for the SEC, AMF, and the main MAR supervisors. The open question is the interesting one. Once the data are clean, will the SEC's shorter T+2 rule still dominate on practical timeliness, or will some European authorities prove that infrastructure and enforcement matter more than a single extra day on paper?

Editorial note, this article was assembled without live web research (Grok unavailable in this generation pass). Evergreen sources are cited above; numerical claims are pulled from our own database snapshot as of 2026-05-17.

Insiders Trades backtest cohort

Our analysis of insider transaction disclosures reveals significant insights into market behavior. In the United States, we observed 187,150 priced trades within the BUY cohort, yielding a mean 90-day return of -2.28% and a win rate of 40.37%. The annualized Sharpe ratio for this cohort stands at -0.09, indicating a lack of risk-adjusted performance. When we expand our view to a global context, encompassing 227,594 BUY filings across 17 markets, the aggregate mean 90-day return is slightly better at -1.70%, with an aggregate Sharpe ratio of -0.08. These figures underscore the challenges investors face when interpreting insider transactions, suggesting that the anticipated benefits may not materialize as expected. It is essential to consider that these findings are based on a specific sample size and time horizon.

Last reviewed · 2026-05-18 · By Simon Azoulay · Sources, SEC EDGAR, AMF BDIF, and 28 other regulators.