WhitmanTrading

Asset Allocation

Asset allocation is the split of a portfolio between broad asset classes, and it decides most of how the portfolio behaves. It is a risk decision rather than a return forecast — the question it answers is how large a fall you can hold through without selling.

Almost every argument about investing is about selection — which fund, which company, which moment. Allocation is the decision underneath all of them, and it is the one that determines how the portfolio behaves in the years that decide outcomes.

How it works

A candlestick chart with a portfolio split across components.
How much of each kind of thing you hold. Illustrative chart - not real market data.

You decide what proportion sits in each broad category — equities, bonds, cash, and whatever else you include — and that proportion, rather than the specific holdings, drives most of what the portfolio does.

The first half of a price series with components behaving differently.
The point is that the parts do not move together. Illustrative chart - not real market data.

The categories are chosen because they behave differently. Holding two things that fall at the same time in the same conditions is not an allocation; it is one position described in two lines.

A section of the price series where a mix dampens a fall.
A mix falls less than its most volatile component. Illustrative chart - not real market data.

Mixing them produces a portfolio that falls less than its worst part and rises less than its best, which is the entire trade being made.

The question it answers

A window of price bars with a deep decline endured.
How large a fall you can hold through. Illustrative chart - not real market data.

An allocation is a statement about how much decline you will tolerate without selling. That is the only version of the question with a checkable answer, because the alternative version — what will perform best — requires knowing the future.

A 100% equity portfolio has historically produced the highest long-run returns and the deepest falls. Adding bonds reduces both. Neither is correct in general; one of them is correct for someone who will sell at the bottom and the other for someone who will not.

And the honest input is what you have actually done, not what you believe you would do. Behaviour in a fall you have lived through is evidence; behaviour in one you have imagined is not.

A worked example

The second half of a price series with two different mixes.
The same market, two different experiences. Illustrative chart - not real market data.

Take a portfolio of 100,000 in a market that falls 40%.

Fully in equities, it falls to 60,000 and needs a 66.67% gain to return to level.

At 60% equities and 40% something that holds its value, it falls to 76,000 and needs 31.58%.

At 40% equities it falls to 84,000 and needs 19.05%.

The market did the same thing in all three. The difference is entirely how much of it you were holding, and the recovery arithmetic is in the drawdown recovery calculator.

Drift

A candlestick series where one component outgrows the others.
The allocation you chose is not the one you have. Illustrative chart - not real market data.

Weights move on their own as prices move, and they move toward whatever has performed best — which means a portfolio quietly concentrates into the thing that has already risen.

A 60/40 portfolio after a strong equity decade is not 60/40 any more, and its behaviour in the next fall reflects the weights it has rather than the ones on the plan.

Correcting it is rebalancing, and doing it with new contributions is cheaper than doing it with sales.

Location as well as allocation

A long-horizon candlestick view with holdings placed in separate accounts.
Which account each part sits in changes the after-tax result. Illustrative chart - not real market data.

Two portfolios with identical allocations can produce different after-tax results depending on which account each holding sits in. Income-generating assets belong inside a wrapper where one is available; assets that generate little until sale can sit outside it.

That is a separate decision from the allocation itself and it is made once, at the start, because correcting it later means selling.

Costs

A candlestick chart annotated with the round-trip cost of a switch.
Every change to the mix is a transaction. Illustrative chart - not real market data.

Changing an allocation costs. On this site’s shared series a round trip measures about 2% of the median bar range of 0.493, and the fee on whatever you hold compounds for as long as you hold it — 75 basis points a year removes 20.2% of a thirty-year pot. The figures are in research/series-measurements.json.

Price bars with planned entries across a stretch.
Which argues for choosing once and leaving it. Illustrative chart - not real market data.

An allocation revised whenever the news changes is not an allocation. Its value comes from being the thing that does not move when everything else does.

The original data

Of the 24,971 unique videos in research/search-study-corpus.jsonl, 4 have a title about asset allocation, at a median of 5,289 views across 4 channels — and 25% use beginner-shaped language. Diversification appears in 3 at 487 and the sixty-forty split in 3 at 959. The counts come from site/rank_investing.py.

A candlestick series with several gaps, the largest of them marked.
A shock moves every component at once. Illustrative chart - not real market data.

Four videos at a 5,289 median, against 449 exchange-traded-fund videos at 12,651. The decision that drives most of a portfolio’s behaviour has a hundredth of the coverage of the decision about which fund to put inside it.

A stretch of price bars cut short at a decision point.
Equities have run for a decade. Increase the weight? Illustrative chart - not real market data.

The answer to the question on that chart is that a decade of strong returns has already increased the weight without your help. Drift did it. Raising the target as well is doing the same thing twice, at the point where the case for it is loudest and the evidence for it is entirely backward looking.

When it fails

The failure is an allocation chosen from what someone believes they can tolerate rather than from what they have demonstrated. It holds perfectly through years when nothing much happens, and it fails in a single week when a fall arrives, because the tolerance was hypothetical. Selling at the bottom converts a paper decline into a permanent loss and usually leads to a more conservative allocation adopted at exactly the wrong moment — so the mistake is paid for twice.

The second failure is treating holding count as diversification. Twenty holdings with one exposure is one bet.

A third is letting drift run unchecked. The risk changes without a decision.

A fourth is ignoring correlations in a crisis. They converge when it matters most.

A fifth is revising the allocation with the news. Its value is in not moving.

And a sixth is separating allocation from location. The same mix in the wrong accounts costs real money.

Diversification is what an allocation is trying to achieve. Risk tolerance is the input the whole decision depends on. And the sixty-forty portfolio is the most common answer to it.

What I actually do

The version of this question that produces a useful answer is not ‘what mix maximises returns’. It is ‘what mix would let me do nothing for two years while the value falls’. Those two questions have different answers, and only the second one is about a decision you actually have to live with.

— Michael Whitman

This page is educational, not financial advice. Test every idea on your own charts before risking money.