A fund lists several exposures. How do I work out what actually moves it?

Not from the weights alone, which is the tempting shortcut. A large regional weight tells you where the money sits, not how sensitive the fund is to that region — a small allocation to something highly sensitive can matter more than a large one to something inert. Two things help more: the index rule, which states what the fund was built to capture, and the fund's own history, which shows what it has actually moved with. Most holdings respond to several things at once, so the realistic output is a short list rather than a single answer — and a short list is enough, because what you are looking for is what recurs across the portfolio.

How many ETFs do I need to be diversified?

For diluting single-company risk, fewer than most people expect: one broad fund already holds hundreds of companies, and the second barely moves that number. For the other kind the answer is not a number at all — what shifts a portfolio's shared exposures is holding something that responds to different conditions, which in practice usually means a different asset class rather than another equity fund. So the honest reply to "how many" is that the count stopped being the useful question somewhere around the second one.

Can two funds with completely different holdings still carry the same risk?

Yes, and it is common. Two funds can share no company at all and still respond to the same conditions — the same sector cycle, the same region, the same sensitivity to interest rates. Overlapping holdings are the visible kind of duplication and the easiest to check; overlapping sensitivities are what decides how the two behave together, and they do not show up in a list of names.

Does this mean one broad fund is better than several?

It does not follow, and it is the easiest wrong turn to take from here. The argument is about what the count measures, not about the merits of owning few things: a single broad fund is one holding that already spreads single-company risk widely and may still be concentrated in a few shared exposures, and several funds may be either. Both properties are set by what sits underneath, and both can be reached from either direction.

Putting it together

More ETFs don't mean more diversification

Counting your funds answers one question about diversification and is silent on another. It tells you something about how exposed you are to any single company failing; it tells you nothing about what your holdings have in common — which is what decides whether they fall together.

7 min de leitura · Atualizado em 2026-09-07

O memo

"Diversified" is doing two jobs. Owning many different things reduces your exposure to any one of them going wrong. It does nothing about whatever they all respond to — and the count speaks only to the first.

Para guardar

Many holdings genuinely dilutes single-company risk. That part of the count is real.
It does nothing about what the holdings share, and that is what moves them together.
Before adding a holding: does it bring an exposure the others do not already have?

6 posições3 condições

sensibilidade alta
sensibilidade baixa
sem sensibilidade

Cada posição responde a mais do que uma condição. Uma coluna responde pelas 6 — e nenhum fundo adicional torna essa coluna mais rala.

A distinção

Diversifying away what is specific to one holding, and diversifying across what your holdings have in common, are two different exercises. The first is largely settled by owning enough different things. The second is not settled by any number.

O erro comum

Reading a long list as evidence of the second kind. The length of the list is real work and real protection against one company failing; it is simply not an answer to the question most people are asking when they ask whether they are diversified.

Em resumo — The useful question is not how many holdings you own. It is how much of the portfolio would be affected by the same turn of events.

Something awkward tends to happen in a bad week. A portfolio spread across eight funds, three providers and several countries falls more or less as one — the same days, the same direction, roughly the same amount. Nothing was wrong with the selection. Each fund is exactly what it says it is.

The week has not shown that the portfolio was badly built. It has shown that the word being used to describe it was answering a different question from the one being asked.

"Diversified" is doing two jobs

Ask whether a portfolio is diversified and you may be asking either of two things, which behave very differently.

The first is how exposed am I to any one holding going wrong — a company failing, a fund's largest position collapsing, a single business turning out to be a fraud. This is what most people picture when they think of not putting eggs in one basket, and it is genuinely addressed by owning many different things. A fund holding several hundred companies has already done most of the work; a portfolio holding several such funds has done essentially all of it. The count really does speak to this.

The second is how much of my portfolio responds to the same conditions — how much of it would be affected together by rates rising, by one region's fortunes turning, by a fall in appetite for equities generally. This is what decides what a bad week looks like. And here the count says nothing at all, because adding more things that respond to the same conditions does not change what the portfolio is exposed to.

Both are diversification. The trouble is that the count is evidence for the first and gets used as evidence for the second.

What holdings actually respond to

An investment does not respond to the name on the fund. It responds to conditions — usually several at once, and to different degrees.

Expectations for economic growth. The level of interest rates. The strength of a currency. The health of one sector. Confidence towards one region. A given fund will have some sensitivity to several of these, strong in places and faint in others, and that profile is what determines how it behaves when conditions shift.

Speaking of "the thing that moves a fund" is therefore a simplification, and a useful one only as far as it goes. What matters for a portfolio is not pinning each holding to one label, but noticing which sensitivities keep recurring across holdings — because those are the ones that will not be diluted by adding another fund that shares them.

Where the concentration hides

Shared exposure survives precisely because the surface looks varied. The companies have nothing in common, so the portfolio appears spread — and in the first sense it is.

A shared exposure	What quietly carries it
Global growth expectations	Most equity funds, whatever the label on them
Interest rates	Long-dated bonds, and the shares whose value rests most on distant profits
The dollar	Assets whose revenues are earned in it, and much emerging-market exposure
One sector's cycle	A sector fund — and any broad index that sector has come to dominate
One country's fortunes	A domestic fund, and the "global" fund heavily weighted towards that country

The last two rows catch people out, because both involve a holding chosen precisely for its breadth. A broad index is only as spread as the weights underneath it, and those weights move with the market rather than with any decision you made — which is a subject of its own.

Reading it for yourself

The visible part is quick. Each fund discloses what it holds, broken down by region, sector and largest positions; an index fund also publishes the rule its index follows. Put two of those side by side and shared names and shared weights show up straight away. Reconciling the index rules behind them takes longer, and is worth the time when two funds look suspiciously alike.

The part that takes judgement is the sensitivity, because it is a question about the world rather than about the document: what conditions would hurt this holding, and do those same conditions appear against the others? Weights help less here than people expect — a large allocation to something that barely moves matters less than a small one to something that moves a great deal — so the index rule and the fund's own history are better guides than the percentage table.

The output is not a score. It is a short list of conditions that keep reappearing, and an honest sense of how much of the portfolio sits behind each.

What this is not an argument for

It is not an argument for owning fewer things, nor for owning more.

Owning many different things is worth doing and it accomplishes something real. What it does not accomplish is the other thing — and no number of holdings ever will, because that property is set by what they respond to rather than by how many of them there are.

So counting your funds is not useless. It is simply not a demonstration that the portfolio is diversified in the sense that matters when everything falls at once.

Confira

Someone holds twenty equity funds from four different providers, spread across several countries, and concludes the portfolio is well diversified. Nothing in that description is false. What has the reasoning established, and what has it not?

Ver a resposta

It has established something real, and it is worth saying first: twenty funds hold a great many companies between them, so no single company failing will do meaningful damage. That kind of diversification is genuinely achieved by owning enough different things, and the count is decent evidence for it. What it has not established is anything about what those twenty have in common. Different providers, different labels and different countries can all sit on top of the same broad exposures — to global growth, to interest rates, to one region's fortunes — and in a week when one of those turns, the twenty can move much as one. So the check is not arithmetic. It is to ask, holding by holding, what conditions would hurt this, and then to notice how much of the portfolio keeps giving the same answer.

Perguntas frequentes

A fund lists several exposures. How do I work out what actually moves it?: Not from the weights alone, which is the tempting shortcut. A large regional weight tells you where the money sits, not how sensitive the fund is to that region — a small allocation to something highly sensitive can matter more than a large one to something inert. Two things help more: the index rule, which states what the fund was built to capture, and the fund's own history, which shows what it has actually moved with. Most holdings respond to several things at once, so the realistic output is a short list rather than a single answer — and a short list is enough, because what you are looking for is what recurs across the portfolio.
How many ETFs do I need to be diversified?: For diluting single-company risk, fewer than most people expect: one broad fund already holds hundreds of companies, and the second barely moves that number. For the other kind the answer is not a number at all — what shifts a portfolio's shared exposures is holding something that responds to different conditions, which in practice usually means a different asset class rather than another equity fund. So the honest reply to "how many" is that the count stopped being the useful question somewhere around the second one.
Can two funds with completely different holdings still carry the same risk?: Yes, and it is common. Two funds can share no company at all and still respond to the same conditions — the same sector cycle, the same region, the same sensitivity to interest rates. Overlapping holdings are the visible kind of duplication and the easiest to check; overlapping sensitivities are what decides how the two behave together, and they do not show up in a list of names.
Does this mean one broad fund is better than several?: It does not follow, and it is the easiest wrong turn to take from here. The argument is about what the count measures, not about the merits of owning few things: a single broad fund is one holding that already spreads single-company risk widely and may still be concentrated in a few shared exposures, and several funds may be either. Both properties are set by what sits underneath, and both can be reached from either direction.