Why does the market only pay for the risk you cannot avoid?

The split from the last lesson is not just a taxonomy. It leads to one of finance's most consequential claims: only one of the two kinds of risk carries an expected reward. The logic is worth more than any model built on top of it.

The argument, in three steps

Step one. Idiosyncratic risk can be removed for free. Unit 3 established this — combine enough genuinely different holdings and company-specific risk largely cancels, at no sacrifice in expected return.

Step two. Nobody pays a premium for something available for free. If an asset offered extra expected return purely to compensate for risk that any diversified holder could eliminate at no cost, diversified investors would buy it until the price rose and the extra return disappeared. The price is set by the buyer who doesn't need the compensation.

Step three. Therefore, in this framework, expected return should compensate only for systematic risk — the exposure nobody escapes by rearranging holdings. Bearing avoidable risk is uncompensated: the risk is real, the reward is not.

What this reframes

It quietly overturns the everyday sentence "riskier investments earn more." The accurate version: investments carrying more of the risk that cannot be diversified away have higher expected returns. Two stocks can show identical total volatility while one carries far more market exposure than the other — and the framework says those two should not have the same expected return.

It also sharpens the concentration argument. An undiversified holder carries the full 35% volatility of a single name but is expected to be compensated only for the systematic 21% inside it. The other 28% is risk carried for free, in the wrong direction — one of the few places in markets where you can pay without buying anything.

The professional caveats

This is a model, and honest treatment includes its limits:

• The formal version (the Capital Asset Pricing Model, and everything built after it) assumes conditions reality only approximates — frictionless trading, shared information, agreement about probabilities.
• Decades of research have documented return patterns the simplest version does not explain, which is why multi-factor models exist at all.
• Real people hold undiversified positions for reasons no model includes: founder stakes, restricted shares, tax consequences, conviction, or information they believe they have.

None of that undoes the core intuition, which has survived every revision: risk you could have removed for free is not the kind markets are in the business of paying for.

Try it now

1. A single name and a broad index are below. Pick one calendar year and, on the single name, find its twenty largest daily moves. Measure the same dates on the index and count how many of them the index also moved more than 1%.

2. That share is roughly the systematic part of the single name's risk — the part the framework expects to be paid for. The rest moved for reasons that belonged to one company, and a portfolio holding fifty such companies would have diversified most of it away without giving up any expected return.
3. Compare the two charts' typical daily ranges as well. The gap between them is, approximately, the piece carried for free. State the idea as description rather than advice: the framework expects compensation for unavoidable risk only, and what anyone does with that observation is their own decision.