Investing Basics
Diversification removes one kind of risk and leaves another standing
Spreading money across many holdings reduces the damage any single one can do, but it cannot reduce the risk that all of them fall together.
By Rosa Iglesias4 min read
Two different risks sharing one word
Risk in investing is usually discussed as a single quantity, which obscures the most useful distinction available. Some risk attaches to a particular holding: a specific company runs into trouble for reasons peculiar to it. Some risk attaches to everything at once: a market falls broadly, and holdings that have nothing in common with each other fall together.
These are conventionally called specific risk and market risk, and they behave completely differently. One of them can be reduced almost to nothing by owning more things. The other cannot be reduced by owning more of the same kind of thing at all, however many of them there are.
Almost every misunderstanding about diversification comes from applying an argument about the first kind to a situation governed by the second.
Why spreading works on the specific kind
The arithmetic is straightforward. If everything is held in one company and that company fails entirely, the loss is total. Spread across a hundred holdings of equal size, the same failure costs one per cent, and the other ninety-nine are unaffected by whatever went wrong.
The reduction happens because the misfortunes are largely unrelated to each other. A supply problem at one firm has no particular bearing on a regulatory decision affecting another, so the events do not arrive together and the good outcomes offset the bad ones. That independence is what does the work, not the number of holdings by itself.
And the benefit arrives quickly. Going from one holding to ten removes a great deal of specific risk; going from ten to a hundred removes noticeably less; going from a hundred to a thousand changes very little. It is a curve with a steep start and a long flat section, which is why the argument for endlessly adding holdings is weaker than it sounds.
Why it does nothing for the other kind
Market risk survives diversification by definition, because it is the portion of movement that all holdings share. Owning every company in a market means owning the market, and if the market falls, everything owned falls with it. There is no number of constituents that escapes this.
This is the risk that actually frightens people, and it is the one that a long list of holdings does least about. It can be reduced by holding things that behave differently from each other — different asset types, different regions, different currencies — but that is a different move from simply holding more of the same, and it comes with its own trade-offs.
It is worth being blunt here: the residual risk is not a rounding error. A well-diversified holding can still fall substantially and can remain below its starting value for extended periods. Diversification changes the character of the risk rather than removing it.
Where diversification is weaker than it looks
Correlations are not fixed. Holdings that behave independently in ordinary conditions have a tendency to move together during severe market stress, which means the protection is at its weakest at the moment it is most wanted. This is a well-documented pattern rather than a theoretical worry, and it is one reason diversification is described as reducing rather than eliminating.
Overlap is a subtler problem. Holding several funds feels like spreading, but if each of them holds many of the same large constituents, the actual exposure is far more concentrated than the number of funds suggests. Looking through to what is held, rather than counting products, is the only way to see this.
Concentration inside a broad index is the same issue arriving from another direction. Where an index is weighted by market value and a small number of constituents have grown very large, a fund tracking it holds a great deal in relatively few names despite containing hundreds.
What it is reasonable to expect from it
Diversification is a way of ensuring that no single mistake, failure or piece of bad luck is decisive. That is a real and valuable property, and it is available cheaply, which is unusual in finance. It is not a way of ensuring that the total does not fall.
There is also a cost side that gets less attention. A widely spread holding cannot outperform substantially either, since the exceptional outcome of any one constituent is diluted by everything else. Giving up that possibility is the price of not being exposed to its opposite, and whether that is a good trade depends on what the money is for.
How much diversification is appropriate, across what, and in what proportions, depends on horizon, on what else a household holds, on tolerance for loss and on rules that vary by country. Investments can fall in value. These are questions for a regulated adviser with the full picture, not for a general description of the mechanism.
Common questions
How many holdings count as diversified?
Most of the reduction in specific risk arrives well before the number gets large, and adding more beyond that point has a diminishing effect. But the count matters less than whether the holdings are genuinely exposed to different things, since many holdings driven by the same factor behave like one.
Does holding several funds diversify better than one?
Only if they hold different things. Several funds tracking similar markets can produce substantial overlap while giving an impression of breadth. The useful check is what is actually held underneath, not how many separate products appear on the statement.
Can diversification lose money?
Yes. It reduces the impact of any single holding failing; it does not protect against a broad decline, and a diversified holding can fall significantly. It also caps the upside from any single exceptional performer, which is the other half of the same trade.
Rosa has written about spending, saving, debt for most of the last decade and is unreasonably interested in the detail nobody else checks.
