The evidence, summarised

A great deal is written about investing and a comparatively small number of things are well supported by evidence.

What is well supported

The findings that recur across studies, markets and periods.

Costs reduce returns predictably. This is arithmetic rather than a finding, and it is the most reliable relationship in the field.

The majority of active funds underperform their benchmarks after costs over long periods, with the proportion rising over longer periods, and persistence of outperformance is weak.

Diversification reduces risk without proportionally reducing expected return, subject to correlations rising during stress.

Investors earn less than the funds they hold, because of the timing of purchases and sales.

More frequent trading is associated with worse net outcomes for retail investors.

Asset allocation explains most of the variation in portfolio outcomes over time.

Time in the market matters more than timing it, since missing a small number of the best days substantially reduces long-run returns.

And defaults and automation improve outcomes more than education or intention.

What is reasonably supported

With more qualification.

Higher starting valuations are associated with lower subsequent long-run returns, over ten years or more, with no short-term timing value.

Factor premiums exist in historical data and have been weaker out of sample, with periods of underperformance longer than most investors tolerate.

Bonds have historically diversified equity risk during growth shocks and not during inflation shocks.

Emerging market growth has not translated reliably into higher equity returns.

And behavioural coaching is a substantial component of the value of financial advice.

What is genuinely uncertain

Stated honestly.

Future returns for any asset class.

The path of interest rates, inflation and growth.

Whether historical risk premiums will persist at historical levels.

Whether the bond-equity relationship will behave as it has.

The appropriate withdrawal rate for a retirement of unknown length in an unknown return environment.

And whether any given active manager has skill.

Anyone claiming certainty about these is claiming more than the evidence supports.

What follows practically

The principles that survive the uncertainty.

Minimise costs, since the effect is certain.

Diversify broadly, since it reduces risk you are not compensated for.

Use tax shelters, since the benefit is certain.

Match the allocation to the horizon and to the capacity for loss.

Automate contributions and increase them with income.

Rebalance on a rule.

Do not act on forecasts.

Do not trade frequently.

Write down the plan and what would change it.

And accept that the outcome is uncertain and that a robust plan is preferable to an optimised one.

What matters more than any of it

Worth stating at the end.

For most households, the financial outcome is determined by income, housing costs, debt, savings rate and pension contribution rate.

Investment selection operates at the margin of these.

Which means the highest-value action for most people reading about investing is increasing the contribution rate rather than refining the portfolio.

And that a mediocre portfolio with a high savings rate beats an excellent portfolio with a low one, reliably.

The behavioural summary

Which is where the largest losses occur.

Selling during declines is the single most damaging behaviour.

Performance chasing is the most common selection error.

Overconfidence produces excessive trading.

Comparison drives participation in things that should be avoided.

And the correctives — automation, a written plan, less frequent checking, less media — are all available and all free.

The honest limits of this

Which should be acknowledged.

Evidence is drawn from particular markets over particular periods, and the future may differ.

Published research is subject to publication bias and to data mining.

General principles do not account for individual circumstances.

And none of it removes the requirement to make decisions under genuine uncertainty.

Which is why a portfolio should be robust to being wrong rather than optimised for being right.

General information only, not investment advice. Investments can fall in value and past performance does not indicate future returns. Consult a regulated financial adviser.