Core and satellite: known limitation 1

Never defining the cap, so the satellite is whatever it has grown into. A holding that trebles has trebled its weight, and a structure with no written limit turns a ten percent allocation into a thirty percent one without a single decision being taken.

Core and satellite: known limitation 2

Using the satellite as a licence rather than a limit. The purpose of a bounded allocation for conviction is that conviction is expensive and should be contained; treating it as permission to trade freely inside the boundary gets the costs of activity without the discipline the structure was supposed to impose.

Core and satellite: known limitation 3

Building a core out of several overlapping funds. A core of five funds that each hold the same large companies is one position wearing five names, at five sets of costs, and it delivers concentration while looking like diversification.

Playbooks / Portfolio

Core and satellite

Hold most of the money in something boring that tracks the market and confine every conviction to a small deliberate remainder, which works only if the remainder is measured, capped and reviewed rather than merely called a satellite.

The rule

Two facts sit awkwardly together. The evidence says that a broad, cheap index fund beats most attempts to improve on it, and people want to act on their views anyway.

Core and satellite is the structure that takes both seriously. Most of the money goes in something boring that tracks a market at low cost and is held through everything. A small, defined remainder is where anything else lives.

It is not a compromise so much as a containment strategy, and understanding it that way is the difference between a structure and a label.

Why the core is the part that matters

The core is doing almost all of the work, and it is doing it by being uninteresting.

The reason to index it is not that active management never works. It is that the identification problem is unsolved: few managers beat a broad index over long periods, and the ones who will do so next decade cannot be picked in advance with any reliability. Studies that track the top performers of one period into the next keep finding them below the index by the end of it.

So the core answers a question nobody has a good answer to by declining to ask it. That is what the three fund portfolio is, and the domicile decision inside it is worth more to most people here than any manager choice, which is fund domicile.

What the satellite is honestly for

Two purposes survive scrutiny.

An exposure the core does not contain. A global equity fund at market weight holds very little of some things, and somebody with a specific and defensible reason to want more of one of them has a real use for a bounded allocation.

Somewhere for conviction to live where it cannot damage the plan. This is the larger of the two, and it is a behavioural function rather than a financial one. A person who will act on a view is better served by a written ten percent than by a resolution not to, because the resolution fails and the ten percent does not.

What the satellite is not for is carrying the return. The honest expectation is that it costs a little, is occasionally interesting, and is small enough that being wrong about it changes nothing important.

The rule that makes it a structure

A satellite that has doubled is no longer a satellite.

This is the whole discipline, and it is where most implementations quietly stop. A ten percent allocation that performs well becomes twenty, then twenty five, and nobody notices because each individual step felt like being right. At that point the portfolio has a concentrated position with an index fund attached, which is a different thing entirely from what was designed.

Rebalancing bands is the mechanism, and it does the uncomfortable thing on purpose: selling the satellite precisely when it has done well. Position sizing decides how large any single bet inside the satellite may be, because a satellite made of one holding is not a risk budget, it is a bet with a nicer name.

The question this audience mostly skips

Everything above describes a fund portfolio, and for most readers of this site the fund portfolio is not the largest holding.

Count the property. An apartment worth one and a half million alongside three hundred thousand in funds is not a core with a satellite. It is a satellite with a rounding error attached, and arranging the smaller part carefully while the larger part sits unexamined is precision applied to the wrong half.

That does not make the property a mistake. It makes it the thing that has to be counted before the structure means anything, which is what concentration limits asks for and what property versus index funds sets out the terms of.

Setting the split

The common shapes run from eighty to ninety percent core, and the precise number matters far less than writing one down before an opportunity arrives, for the same reason every other limit on this site is written in advance.

Then say which category each holding belongs to, out loud. Most people doing this exercise for the first time discover that several holdings they thought of as core are in fact satellites, and that the satellite total is already well past whatever they would have chosen.

The arithmetic

The structure Core broad, cheap, index tracking, held through everything Satellite a small, capped share for specific bets A common shape Core 80 to 90 percent Satellite 10 to 20 percent The precise split matters far less than writing it down. What the core is for Getting the market return at low cost, reliably, without a manager selection decision. What the satellite is for Two honest purposes. 1. An exposure the core does not contain. 2. Somewhere for conviction to live that cannot damage the plan. The rule that makes it work A satellite that has doubled is no longer a satellite. Rebalance it back or admit the structure has been abandoned. The question this audience skips Count the property. A 1.5m apartment beside a 300k fund portfolio is not a core with a satellite. It is a satellite with a rounding error attached.

Where it breaks

Never defining the cap, so the satellite is whatever it has grown into. A holding that trebles has trebled its weight, and a structure with no written limit turns a ten percent allocation into a thirty percent one without a single decision being taken.
Using the satellite as a licence rather than a limit. The purpose of a bounded allocation for conviction is that conviction is expensive and should be contained; treating it as permission to trade freely inside the boundary gets the costs of activity without the discipline the structure was supposed to impose.
Building a core out of several overlapping funds. A core of five funds that each hold the same large companies is one position wearing five names, at five sets of costs, and it delivers concentration while looking like diversification.
Counting only liquid assets when deciding the split. For most people reading this the largest holding is a property, and a fund portfolio arranged into a careful core and satellite beside an unexamined apartment is precision applied to the smaller half of the problem.
Assuming the satellite is where the return comes from. The evidence on manager selection is that few beat a broad index consistently and fewer can be identified in advance, so the honest expectation for a satellite is that it will cost something and occasionally be interesting, not that it will carry the portfolio.

When to use it

When somebody wants to hold an index fund and still act on a view, which is most people. It is also the right structure for anyone who has accumulated holdings one at a time and needs a way to describe what they now own without selling all of it.

Sources

Last reviewed 28 August 2026. Educational research, not personal advice. Disclosure standards.