Asset Allocation: What the Decision Is

Scope — and this article carries the pillar's refusal in its own body rather than relying on the hub. MarketClue does not tell any reader what their allocation should be, does not publish age-based formulas or rules of thumb, does not offer model portfolios, and operates no questionnaire that terminates in a set of weights. This article explains what the allocation decision consists of and how much it determines — nothing below is a suggestion. All figures use the illustrative teaching parameters fixed on this pillar's hub and are not forecasts of any market.

It is also the decision most often described using a statistic that does not mean what it is quoted as meaning, so this article deals with that first.

The most misquoted number in the subject

The claim usually encountered is that asset allocation explains around 90% of investment returns. It is a real finding, it is widely repeated, and the thing it explains is not returns.

A study published in the Financial Analysts Journal in 2000 separated three different questions and answered each. Examining ten years of monthly returns for 94 balanced mutual funds along with pension fund data, it found:

Worked example — the difference between the 90 and the 40 is the whole point. The 90% figure says that when a given portfolio's returns move around from period to period, almost all of that movement traces to what it holds in broad categories — which is unsurprising, since a portfolio that is mostly equities will move when equities move. The 40% figure answers the question most readers actually have: if one portfolio does better than another, how much of the difference is the allocation? The study puts it concretely: where one fund returns 13% and another 8%, on average about 2 points of the 5-point difference is explained by the difference in asset mix, and the remaining 3 points by other factors including selection, timing and fees. So allocation explains most of why a portfolio moves and a minority of why it beat or lagged another one. Quoting the 90% as though it answered the second question is the commonest error in this subject, and it is usually made by people arguing that nothing else matters.

What the decision actually spans

On the hub parameters, taking $10,000 held for twenty years at the expected returns implied by each mix.

The extremes differ by a factor of 2.11. That is what the decision is worth in expectation, before anything is known about which securities are held.

And what it does to the range of outcomes

Expected values conceal the more important half. Simulating 200,000 twenty-year paths on the same parameters:

Worked example — the middle rises and the bottom falls, simultaneously. Moving from 20% to 100% in equities lifts the median outcome from $29,840 to $45,418 — and lowers the fifth-percentile outcome from $19,149 to $14,477. More of the higher-returning asset improves the typical result and worsens the poor one. The spread between the best and worst cases widens from 2.4 times to 9.2 times. This is what an allocation decision is: not a choice of how much to earn, but a choice about which distribution of outcomes to stand inside. No point on this table is better than any other, because which one is preferable depends on how much the poor outcome matters relative to the good one — a question about a person's circumstances and obligations, not about markets, and one this portal does not attempt to answer for anyone. (Simulated on annual returns drawn from the stated means and standard deviations; the percentiles vary by a few hundred dollars from one run to another, and the shape — median up, tail down, spread widening — does not.)

What the decision requires, and why nobody else can make it

The inputs are personal and none of them is observable from outside. When the money is needed — a horizon of three years and one of thirty are different problems. What else the person has — income stability, other assets, obligations. What a poor outcome would cost them — an inconvenience or a serious problem. And what they will actually do when the portfolio falls, since an allocation abandoned at the worst moment delivers neither its expected return nor its expected risk.

Those four are exactly what a formula cannot contain, which is why this portal treats age-based rules of thumb as a category error rather than as a simplification. An allocation derived from a single variable answers a question with four.