Asset Class

An asset class is a set of investments grouped together because they share the same economic characteristics and are driven by the same things, so that members of the group tend to move together and to move differently from members of other groups. The SEC's investor glossary entry is one sentence long: asset classes are "investments that have similar characteristics," and "the three main asset classes are stocks, bonds, and cash."

The brevity of that entry is itself informative. The classification is a working convention rather than a formal taxonomy: it exists because grouping investments by behavior makes a portfolio's risk describable in a few numbers. What the classes are called, and how many of them there are, varies by whoever is drawing the picture. That is not a defect to be corrected so much as a fact to be handled, and handling it starts with knowing that the boundaries are drawn rather than discovered.

What actually makes a grouping a class is shared drivers, not shared paperwork. Two investments belong in the same class when the same conditions help or hurt both of them: corporate profits and the price investors will pay for them, in one case; the level of interest rates and the creditworthiness of borrowers, in another. The practical consequence follows directly. Holding twenty things from one class is not the same as holding twenty things, because a shock that reaches the driver reaches all twenty at once. That is the starting point for diversification, which takes the question up from there.

Where the boundaries sit is genuinely disputed, and the disputes are not academic. Real estate is the clearest case. Property is often treated as its own class, while a real estate investment trust is a company whose shares trade on an exchange, and classifying the same holding one way or the other moves a real percentage in a real portfolio. Commodities are grouped together by the fact that they are physical inputs rather than by any shared driver, and oil and gold have very little in common economically. Crypto assets are argued about on exactly these terms. Cash sits inside the conventional three even though its role, holding value for near-term spending, is different in kind from the role the other two play. And "alternatives" is not a class at all in the sense used here, since it is defined by what its members are not.

The most useful discipline is to ask what a grouping predicts. A classification earns its place if knowing which class something belongs to tells you something about how it will behave. By that test, splitting a portfolio into US large-company stocks and US large-company value stocks creates two labels and one exposure, because the same conditions drive both. The reverse also holds: two holdings that look unrelated on a statement can turn out to be one bet if the same conditions drive them.

This is why "a new asset class" deserves a question rather than a nod. The phrase is doing real work when a product is genuinely driven by something the existing groups are not, and no work at all when the product is a repackaging of holdings that already sit inside a class. Neither the phrase itself nor a fund's own category label settles it. What settles it is what the thing responds to, and that can be asked of any product in one sentence.

Decide what the classes are, sort every holding into one of them, and add up the totals. The percentages that come out are the portfolio's allocation, and every later judgment about risk rests on them. Because the sorting step is a judgment, the same portfolio can produce different pictures depending on how a few holdings are treated, which is why the classification should be written down once and applied consistently.

A hypothetical example of the classification changing the answer without changing the money. Theo's portfolio totals $500,000: a US stock index fund holding $300,000, a bond fund holding $150,000, and a real estate investment trust fund holding $50,000.

Treating real estate investment trusts as stocks, because they are shares in listed companies, the portfolio is $350,000 in stocks and $150,000 in bonds, which is 70% and 30%. Treating real estate as its own class, it is 60% stocks, 30% bonds and 10% real estate. Every holding is identical in both versions. A target of "no more than 65% in stocks" is breached in the first version and comfortably met in the second, and neither classification is wrong. What matters is that the same one is used when the target is set and when compliance with it is checked.

Pros

• Compresses a portfolio of many holdings into a few numbers that describe how it is likely to behave.
• Makes different portfolios comparable, since the categories are broadly shared even where the edges are not.
• Sorting by shared drivers exposes holdings that look distinct but respond to the same conditions.
• Gives a household a vocabulary for setting and checking a target mix.

Cons

• The boundaries are conventions, so two classifications of the same portfolio can produce materially different percentages.
• Grouping by label rather than by driver can hide an exposure, since two holdings in different classes may still rise and fall together.
• The conventional classes are broad enough to contain very different things, as a Treasury bill and a long low-rated corporate bond both sitting under bonds.
• A category label attached by a fund's marketing is not evidence about what drives the fund's returns.