What does not workSAVNG Research · October 7, 2026Pre-registered

Insider selling does not warn that a company will fail

The claim. Comparing companies only with others of the same size, the ones whose CEO or chief financial officer sold shares in the past year went on to fail slightly less often (0.19% within a year) than same-size companies where they did not (0.28%). The old pattern, that executives at doomed companies sell less, was mostly company size.

What happened and what changed

Data: 2.6 million open-market insider sales reported on Form 4 from 2006 to 2026, each used only from the day it was filed, with the seller's role (chief executive, finance chief, other officer, director, 10% owner).
Failure test: about 175,000 company-quarters from 2011 to 2024 (6,114 companies, 1,014 quarters followed by a bankruptcy or delisting for cause). Each company was compared only with companies in the same tenth by market value in the same quarter, because small companies both fail more and have fewer insiders selling.
Rules set before looking: selling counts as a warning if same-size companies with selling fail at least 1.5 times as often, with a reliable difference; as a sign of safety if they fail at most two-thirds as often.

In the previous 12 months	Failed within a year	Same-size companies without it	Ratio	Reading
CEO or finance chief sold shares	0.19%	0.28%	0.69	Slightly less failure, not more
Any insider sold shares	0.38%	0.35%	1.07	No difference
CEO or finance chief sold more than 10% of their stake	0.12%	0.23%	0.55	Less failure, not more
No insider buying or selling at all	1.26%	1.21%	1.04	No difference

Neither pre-set rule passed. The "less failure" reading for chief-executive and finance-chief selling was 0.51 in 2009–2016 and 0.81 in 2017–2024, fading toward no difference.

Where the old pattern came from: 16% of quarters before a failure had a CEO or finance-chief sale, against 48% of all surviving companies. Against survivors of the same size it is 16% against 21%. Most of the gap was size: bigger companies have more executives selling stock they were paid in.
Silence: companies where no insider traded at all looked 2.6 times as likely to fail. Matched for size, the difference vanishes (1.26% against 1.21%).

Bought the day after a CEO or finance-chief sale, held 3 months, against same-size companies with no insider selling	2006–2016	2017–2021	2022–2026
Average, counting each company once	−2.8 pts	−1.1 pts	−4.8 pts
Average, counting every sale	+0.0 pts	+2.1 pts	−1.6 pts

The two ways of counting disagree because companies whose executives sell rarely lagged, while large companies whose executives sell all the time did as well or better. Which kind a company is only becomes clear afterwards, so this is not a trade.

Reproduce: research-engine/strategies/is01_insider_selling.py → C:/FirstRate/work/out/is01/card.md

Insider buying, which does carry information →

Where it does not apply

Selling does not predict weak returns you could trade on either: counted company by company, sellers trailed same-size non-sellers, but counted trade by trade or month by month the gap disappears or reverses, because the lag sits in companies whose executives rarely sell. The 10b5-1 plan flag only exists from 2023, so most older sales cannot be split into planned and discretionary.

SAVNG is an educational research publication, not a Registered Investment Adviser. Everything here is general-circulation information produced by an automated pipeline — the same for every reader — and is not personalized investment advice, an offer, or a recommendation to buy or sell any security. Any performance figure is hypothetical; past performance does not predict future results. Details.

Published by SAVNG Research. Every figure on this page is read from the study data at render time; if the study changes, the page changes with it.

← All research