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.
Tax-loss selling has a calendar, and the calendar is annoyingly local.
Tax-loss selling is one of those market habits that survives because it is partly rational, partly institutional, and partly a triumph of deadlines over theory. Investors do not wake up on 28 December with a fresh view on discounted cash flow. They wake up with realised gains, unrealised losses, and a tax adviser who has become suddenly popular.
For insider transactions, the story is more complicated. Executives and directors are not ordinary investors. Their trades are constrained by disclosure rules, closed periods, internal dealing codes, reputational concerns, and in some cases the small inconvenience of actually knowing things. That makes seasonal selling patterns by jurisdiction less obvious than the standard retail tax-loss narrative. It also makes them more interesting.
This article looks at month-end and year-end clustering in insider sells across France, the United States, Germany, and the United Kingdom. The short version is that the calendar effect should exist, but it should not look the same everywhere. The tax year differs. The treatment of capital losses differs. The wash-sale or bed-and-breakfast style anti-avoidance rules differ. The disclosure architecture differs. If one expects a single global December spike, one is importing an American habit into markets that did not ask for it.
Seasonality in selling is not a mystery. Taxes create a hard stop. If losses can offset gains within a tax year, the option value of waiting collapses as that year closes. The result is bunching. The exact date of that bunching depends on the legal year-end, the ability to carry losses forward or backward, anti-repurchase rules, and whether the investor can plausibly transact at all.
For ordinary investors, the classic mechanism is simple. Realise losses before the tax year closes, use them to offset gains, then either stay out of the name or re-enter after the anti-avoidance window. For insiders, add several layers:
In the EU and UK, persons discharging managerial responsibilities, PDMRs, face restrictions around financial-reporting periods. Under MAR, a closed period of 30 calendar days applies before publication of annual and interim financial reports, subject to narrow exceptions. That can push legitimate sales away from exactly the period when tax incentives are strongest. A December-heavy tax motive can therefore collide with a late-year blackout.
In the US, there is no MAR-style harmonised closed-period regime imposed by statute across issuers, but issuers commonly maintain insider-trading policies and blackout windows around earnings. Rule 10b5-1 plans also matter, because they can pre-schedule trades and reduce visible discretion. A December cluster in US insider sales may therefore be partly tax, partly plan execution, and partly compensation housekeeping.
A large share of insider sales, especially in the US, are linked to equity compensation. Option exercises paired with sales, tax-withholding disposals, and automatic sales under plans are not the same as a discretionary decision to crystallise a capital loss. If one simply counts all sales, one is likely to confuse payroll plumbing with tax behaviour.
The US tax year aligns with the calendar year. The UK tax year ends on 5 April. France and Germany use calendar-year taxation for individuals, but their capital-gains regimes and offset mechanics differ from the US template. That means the same economic motive can show up in different months, with different intensity, and in different investor populations.
If one had to pick a single market where tax-loss selling should be most visible, the US would be the obvious candidate. The tax year ends on 31 December. Capital losses can offset capital gains, and then a limited amount of ordinary income, with unused losses carried forward. The wash-sale rule under Internal Revenue Code Section 1091 discourages immediate repurchase within 30 days before or after the sale if the taxpayer wants to preserve the loss deduction. The behavioural result is familiar: sell losers into year-end, avoid the wash-sale trap, revisit in January.
US insiders report most transactions on Form 4 within two business days. That gives researchers decent visibility. It does not, however, guarantee clean interpretation. A December sale by a chief executive could be:
The SEC has spent years refining the disclosure around Rule 10b5-1 plans, precisely because pre-arranged trades can otherwise masquerade as informed timing. For seasonality work, those plan-related flags are useful. They do not eliminate noise, but they reduce the chance of attributing every December sale to tax motives and every January lull to virtue.
The broader market literature has long documented tax-loss selling pressure in US equities, especially among retail-heavy, loser stocks, with some reversal in January. The insider-specific literature is thinner and more mixed, because insiders are a narrow and institutionally constrained subset. Still, the mechanism is plausible where insiders hold meaningful unrestricted stock outside formal plans.
A practical expectation for US insider data is therefore not merely a December spike in all sales, but a stronger concentration in discretionary open-market sells of underperforming names, especially in the second half of December, followed by some normalisation in January once the wash-sale window and year-end reporting pressure pass.
Month-end clustering is a different animal. Some of it is administrative. Boards meet, windows open after earnings, payroll and vesting cycles hit, and treasury departments prefer tidy dates. Tax-loss logic can reinforce month-end in December, but month-end patterns in the US are often too broad to be called tax effects without further controls.
A useful test would compare the final five trading days of December with the final five trading days of other months, stratified by prior stock underperformance and by transaction type. Without that, one is mostly measuring office habits.
The UK is where lazy cross-border comparisons go to die. The tax year for individuals ends on 5 April, not 31 December. If tax-loss selling is materially shaping insider behaviour, a UK seasonal pattern should lean toward March and early April, not necessarily December.
For UK taxpayers subject to capital gains tax, losses are generally available to offset gains, with specific rules on claiming and carrying forward. The anti-avoidance framework also differs from the US. The well-known share matching rules, including the same-day rule and the 30-day “bed and breakfast” rule, affect how repurchases are treated. That changes the tactical shape of loss harvesting.
The implication is straightforward. A UK insider who wants to realise a loss for tax purposes has little reason to care deeply about 31 December unless another constraint, such as bonus timing or liquidity needs, intrudes. The more relevant deadline is the tax-year end in April.
Post-Brexit, the UK retained a domestic version of MAR. PDMR transactions must still be disclosed promptly, and the 30-day closed-period framework remains central. That creates an awkward interaction with the tax year because many UK companies report annual results in spring, exactly when tax incentives intensify. In other words, the tax motive and the dealing restriction can collide head-on.
This is one reason UK insider tax-loss selling may appear less cleanly bunched than one might expect from tax rules alone. A director who would like to sell in late March may simply be in a closed period. The sale may then be pulled forward, delayed, or not happen at all.
For UK data, the sensible hypothesis is a relative increase in discretionary sells in the weeks before 5 April, especially among names with poor trailing performance, but with distortion from reporting blackouts. December may still show activity, but it should not dominate on tax grounds alone.
That is a useful reminder that “year-end” is not a universal concept. In markets, it usually means “the date after which your accountant becomes less forgiving”.
| Market | Regulator | Rule | Deadline | Notes |
|---|---|---|---|---|
| FR | AMF | MAR Art 19 | T+3 | Calendar-year tax system, EU-style PDMR disclosure and 30-day closed periods before annual and interim reports. |
| US | SEC | Section 16 Form 4 | T+2 | Calendar-year tax system, wash-sale rule, issuer blackout policies and 10b5-1 plans shape observed timing. |
| DE | BaFin | MAR Art 19 | T+3 | Calendar-year tax system, EU MAR framework, domestic tax treatment of losses and withholding details affect incentives. |
| UK | FCA | UK MAR Art 19 | T+3 | Tax year ends 5 April, so tax-loss incentives may cluster in March and early April rather than December. |
France and Germany belong together for one reason and apart for another. They share the broad EU market-abuse architecture for insider dealing disclosures. They do not share identical tax systems, investor habits, or ownership structures.
In France, PDMR transactions are disclosed under MAR and overseen by the AMF. The tax year is the calendar year, which means a December tax-loss motive is at least directionally plausible. But French insider trading data often reflect concentrated ownership, family control, and governance structures that differ from the more dispersed US model. That matters because insider sales may be less about annual portfolio management and more about strategic liquidity, succession, or holding-company financing.
French taxation of capital gains and losses for individuals has changed over time, including the flat tax regime and the treatment of offsets under different elections. Those details affect incentives, but they do not erase the basic point that losses are more valuable before year-end than after it.
The practical expectation is modest December bunching in discretionary sales of poor performers, but likely weaker than in the US because the insider base is less retail-like in its behaviour and because closed periods around annual reporting can interfere.
Germany also uses the calendar year for tax purposes and applies MAR disclosure rules through BaFin’s supervisory framework. Again, that makes a late-year tax motive plausible. Yet Germany’s market structure and executive ownership patterns differ from the US. There is less reason to expect a dramatic, retail-style year-end flush in insider transactions.
Another issue is that German insider disclosures can be relatively sparse in some names and episodic in others, particularly where controlling shareholders dominate. A seasonal pattern may therefore be visible only after controlling for issuer size, free float, and the role of founder or family insiders.
Even under the same broad EU dealing rules, tax-loss selling can differ because the economic owner differs. An executive with substantial unrestricted stock and active tax planning behaves differently from a supervisory board member with occasional grants or a family holding company managing a strategic stake. The filing may look similar. The motive does not.
This is where a jurisdictional article can become too neat for its own good. France and Germany may both show “December-ish” selling pressure. The underlying drivers can still be materially different.
Year-end gets the headlines because taxes are visible and December is cinematic. Month-end is more mundane. That does not make it unimportant.
Insider transactions often cluster around predictable corporate events. Earnings releases open dealing windows. Board meetings happen on fixed calendars. Vesting and payroll cycles are monthly or quarterly. Brokers and legal teams prefer standard processing windows. If one sees elevated selling around month-end across all four jurisdictions, the default explanation should be administration, not tax.
The interaction to watch is not a generic month-end effect. It is whether the final month of the relevant tax year shows a stronger last-week concentration than other months. In the US, that means December. In the UK, March and the first days of April. In France and Germany, again December.
A robust design would compare:
If underperformers are disproportionately sold into the tax-year-end window, the tax-loss story gains credibility. If all names show the same bunching, one is probably looking at process rather than tax.
Tax-loss selling requires losses. That sounds obvious because it is obvious. Yet many seasonal studies forget to isolate names with unrealised losses relative to plausible acquisition bases. For insiders, cost basis is rarely public, so researchers use stock underperformance as a proxy. It is imperfect but useful. A December cluster in sales of recent winners is not tax-loss selling. It is something else, possibly the oldest something else in finance, namely taking money off the table.
The bad version of this article would stop at legal calendars and call it insight. The better version asks what a credible empirical test would require.
You need to separate at least six buckets:
Only the first bucket is a clean candidate for tax-loss timing. The rest are mostly contamination for this question.
This is the central cross-jurisdiction fix. If you align the UK on December, you are asking the wrong question politely. Build event time around each jurisdiction’s relevant tax-year end and compare behaviour in the final 30, 15, and 5 trading days.
A sale cannot happen during a closed period unless an exception applies. That means the opportunity set varies by issuer and month. A proper model should include issuer reporting calendars, or at least annual and interim report months, to avoid treating forbidden days as days of low desire.
Because cost basis is unavailable, use trailing returns and distance from 52-week high as rough indicators. The tax-loss hypothesis predicts stronger end-of-tax-year selling in names that have performed poorly.
If tax-loss selling is active, one might also expect a relative suppression of insider buys in the tax-year-end window for underperformers, followed by some rebound after anti-avoidance windows pass or after reporting windows reopen. The absence of any asymmetry between buys and sells weakens the tax explanation.
Founder-led firms, state-influenced issuers, family-controlled groups, and widely held corporates generate very different insider-flow patterns. Country effects can be ownership effects wearing a flag.
The purpose of the closed period is to prevent trading by persons discharging managerial responsibilities during a period when they are more likely to possess inside information.
Without the internal Sigma extract for this article, the honest position is to state priors rather than pretend to have measured them. Happily, honesty is still cheaper than a correction.
Expected pattern: the strongest tax-year-end clustering in discretionary sells, concentrated in December and strongest in underperforming names.
Why: calendar-year taxation, established tax-loss harvesting behaviour, transparent filing regime, and a deep population of insiders with meaningful personal equity exposure.
Main confounders: 10b5-1 plans, compensation-related sales, issuer blackout policies.
Expected pattern: weaker December effect, stronger clustering around March and early April, subject to distortion from annual-report closed periods.
Why: tax year ends on 5 April, share matching rules shape repurchase tactics, UK MAR preserves PDMR disclosure and blackout structure.
Main confounders: collisions between tax incentives and spring reporting blackouts, concentrated holdings in some UK issuers.
Expected pattern: some December concentration in discretionary sells, but likely moderate rather than dramatic.
Why: calendar-year taxes support year-end realisation, but insider ownership structures and strategic holdings may reduce the share of trades driven by annual tax housekeeping.
Main confounders: family and control structures, lower frequency of purely discretionary trading, MAR closed periods.
Expected pattern: similar directional tendency to France, possibly with even more heterogeneity by issuer type.
Why: calendar-year taxation and MAR support a year-end effect in principle, but ownership structure and lower discretionary trading intensity may mute it.
Main confounders: controlling shareholders, episodic filings, strategic rather than portfolio-driven disposals.
Tax-loss selling in insider data is not merely a compliance curiosity. It can matter for signal extraction.
If a market has strong tax-year-end clustering in insider sells, those sells should be less informative on average than off-season sells. A December US sale in a beaten-up stock may say less about management’s private view than a May sale in the same name. Conversely, a UK insider sale in late March may deserve the same discounting treatment that US analysts already apply to late-December disposals.
That matters for any strategy or editorial framework that treats insider selling as a bearish signal. The signal is conditional. Calendar effects can lower its information content. Ignoring that is a good way to confuse tax administration with alpha.
The practical next step is plain enough: build a jurisdiction-aware seasonality panel using only discretionary open-market sells, align each market to its actual tax-year end, and test whether underperformers show excess selling in the final trading days once blackout windows and plan-related trades are stripped out. If that excess survives, we have a real seasonal effect. If not, the open question is even better, namely whether most apparent tax-loss clustering in insider data is just corporate process dressed up as behavioural finance.
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.
Our analysis of the UK market reveals a notable absence of priced trades among insiders, resulting in a mean 90-day return that remains unreported. This lack of data underscores the challenges in assessing tax-loss selling patterns in this region. In contrast, when examining the broader global universe, which encompasses 227,594 BUY filings across 17 markets, we observe an aggregate mean 90-day return of -1.70% and an annualized Sharpe ratio of -0.08. These figures suggest that, on average, insider buying does not yield positive returns in the short term, potentially indicating a pattern where tax-loss selling may not be as prevalent or effective as theorized. However, it is crucial to note that the sample size and time horizon for the UK data limit the robustness of these findings.
Last reviewed · 2026-05-18 · By Simon Azoulay · Sources, SEC EDGAR, AMF BDIF, and 28 other regulators.
Learn how to construct a beta-neutral insider portfolio to reduce drawdowns and improve stock selection without broad ma...
Learn how to optimize insider trading backtests using a strict walk-forward protocol to avoid hindsight bias and ensure ...
This article explores the implications of multiple-testing inflation in quant research, particularly in insider trading ...
Large insider trades can distort scoring models, often driven by factors unrelated to fresh information. A cap on trades...
Explore the relationship between insider trading and sector ETF rebalances, particularly in technology and energy sector...
Insider transaction disclosures can significantly influence stock prices. This article analyzes the abnormal returns fro...
