Behaviour
Behavioural finance has a replication problem
Some of the most quoted findings in the field have held up poorly under retesting. Knowing which parts are solid matters more than collecting the anecdotes.
Everything below about the evidence behind behavioural finance comes from what actually happens rather than from what is supposed to.
What holds up in practice
- Several widely repeated psychology findings have failed or weakened on retesting.
- Documented market anomalies have often shrunk after publication.
- The core behavioural patterns relevant to investors are among the better supported.
What the replication issue is
Across psychology and adjacent fields, systematic attempts to repeat well-known experiments have often produced weaker effects than the originals, and sometimes none. The causes are structural rather than scandalous: small samples, flexible analysis choices, and a strong bias toward publishing striking results. This has prompted genuine methodological reform, which is a sign of a field working rather than failing.
It does mean that any single memorable finding deserves less weight than its popular reputation suggests.
Popular accounts amplify the weakest parts
Books and articles select for surprising, tidy results with a clear story, which is exactly the profile most vulnerable to being overstated. Precise numbers attached to psychological effects are particularly suspect, since magnitudes vary substantially by context.
For most people, a finding repeated for two decades in popular writing has often not been retested seriously in that time. Repetition and evidence are easy to confuse and are not related.
Anomalies fade after publication
Market patterns documented in academic work have frequently become weaker in the data that followed publication. Some of that is markets adapting once a pattern is known, and some is that the original finding was partly a product of the data it was found in. Either explanation argues against building a strategy on a published anomaly.
For most people, it also explains why so many strategies work impressively until they are adopted.
What holds up reasonably well
That people react more strongly to losses than to equivalent gains is well supported as a direction, if not as a fixed ratio. That individual investors who trade more have tended to do worse net of costs has been found in several independent datasets. That defaults strongly influence participation has been replicated across multiple countries and systems.
That fund performance persists weakly and cost predicts relative performance better than past returns is supported by repeated analysis over long periods.
How to hold the uncertain parts
Treat behavioural claims as descriptions of tendencies with uncertain magnitudes rather than as laws with coefficients. Prefer conclusions that are robust to the size of the effect, since those survive if the number turns out smaller. Every practical recommendation in this area follows from direction alone: check less, write things down, automate, keep costs low.
None of them require any specific number to be correct.
If that does not fit your week, it is not a failure of willpower.
Why this matters for you
A reader who learns behavioural finance as a list of quirky biases will mostly acquire vocabulary for explaining past mistakes. A reader who learns the mechanisms can design a process that does not depend on being unusually disciplined.
Scepticism about the evidence base and confidence in the practical conclusions are compatible positions here. It is also a reasonable filter for any source that quotes precise psychological numbers without qualification.
The takeaway
Trust the directions, distrust the numbers, and act on the conclusions that survive either way.
The version you keep doing is the version that works.
Questions readers ask
Does this mean behavioural finance is unreliable?
It means the field contains both well-supported patterns and popular findings that have not held up. The practical conclusions for investors rest mostly on the better-supported end.
How should I read a striking behavioural claim?
Ask whether the conclusion changes if the effect is half the stated size. If it does, the claim is doing more work than the evidence supports.
