What Is a Risk Premium?
Risk premium is the additional expected return an investor requires for holding a risky asset rather than a government bill. It is an expectation rather than a guarantee, so a premium can be real and still fail to appear over any particular holding period.
A risk premium is the extra return you expect for accepting uncertainty. Everything difficult about the concept follows from the word expect.
How it works
A government bill offers a near-certain return. Anything with uncertain outcomes must offer more, or nobody rational would hold it.
The difference is the premium. It is priced into the asset at purchase, which means it is a property of the price rather than a property of the asset.
And it is an average across possible outcomes. Any single path can deliver less than a government bill returns, or a loss, for a very long time.
Why it has to be able to fail
A premium that always paid would not be a premium. It would be a free lunch, competed away by anybody able to borrow cheaply and buy the asset.
The possibility of it not paying is the mechanism. Investors demand the extra return precisely because the bad outcomes are real, and removing them would remove the compensation.
Which is why long periods of underperformance are consistent with a premium existing, not evidence against it — an uncomfortable position because it makes the claim hard to falsify.
A worked example
The equity risk premium is commonly estimated at a few percent a year over government bonds, from long historical series.
Those series contain multi-decade periods where equities underperformed bonds. Investors holding through them experienced the risk with none of the premium, over horizons longer than most working lives.
And the historical estimate is itself uncertain. Survivorship matters — the markets with the longest continuous records are the ones that were not closed by war or revolution, which biases the sample upward.
None of that means the premium is fictional. It means the number is less precise than its usual presentation, and the horizon required for it to appear is longer than most people plan for.
Forward-looking versus historical estimates
Historical premia are what happened. They average realised returns over a past period and assume the future resembles it.
Forward-looking premia are implied from current prices. Take today’s valuations and an assumption about future cash flows, and back out the return being priced.
The two frequently disagree. After a long bull market the historical premium is large and the implied forward premium is small, because high prices mechanically imply lower future returns.
And the implied version is more useful for a decision. It describes what you are being offered now rather than what somebody was offered decades ago, though it depends entirely on the cash flow assumptions fed into it.
The original data
On this site’s shared series 95% of bars sit below a prior peak, the maximum decline is 3.76%, the longest below-peak stretch runs 73 bars, and that stretch finished +3.61%.
That last pair of figures is the concept in miniature. Seventy-three consecutive bars below a prior peak, ending higher than it started — the reward was present and required sitting through a long stretch where it was not visible.
And this site’s fee measurement is the deduction from it: 75 basis points costs 20.2% of a thirty-year balance, 150 costs 36.5%. A premium of a few percent, less a fee of that size, is a materially smaller premium — and the fee is certain while the premium is not.
The other premia people try to harvest
Term premium, for lending longer rather than shorter, which has been negative for extended periods.
Credit premium, for lending to riskier borrowers, most of which compensates for defaults that do eventually arrive.
Illiquidity premium, for holding something hard to sell — genuinely real and frequently overstated, because the measured returns of illiquid assets are smoothed by infrequent valuation.
And the various factor premia — value, size, momentum and others — where the honest position is that publication and product launch tend to shrink whatever was documented, and several have underperformed substantially since the papers describing them appeared.
Why the measured number keeps shrinking
Estimates have fallen over time. Early studies of the equity premium produced figures well above what most current work suggests, and the revisions have been consistently downward.
Survivorship explains part of it. The markets with long usable records are the ones that were not interrupted by war, expropriation or closure, and including the interrupted ones lowers the average.
Valuation levels explain another part. Much of the historical return came from valuations rising, which is a one-time repricing rather than a repeatable premium.
And the honest conclusion is a range rather than a number. Anybody quoting the equity premium to one decimal place is reporting a point estimate from a distribution wide enough to include zero over many plausible horizons.
When it fails
The characteristic failure is treating a premium as an entitlement. An investor reads that equities earn a few percent over bonds, plans around it, and holds through a decade where they do not. The premium was not withdrawn and nothing was miscalculated — it is an average across a distribution of outcomes, and this particular decade drew from the unfavourable part. Planning on an expected value as though it were a schedule converts a reasonable long-run assumption into a specific short-run promise that nothing was ever offering.
A second failure is using a historical estimate at a market peak, where implied forward returns are lowest.
A third is ignoring survivorship in the long series the estimates are drawn from.
A fourth is chasing a documented factor premium after publication has already attracted the capital.
And a fifth is forgetting that fees come out of it. A premium of 4% less a 1% fee is a 3% premium, and that quarter was certain from the start.
Related
Market risk covers the exposure the largest premium compensates. Variance risk premium covers one of the more reliably measured ones. And tail risk covers the outcomes a premium is ultimately payment for.
The single most important sentence about risk premia is that they are not owed to you. A premium is compensation for the possibility of a bad outcome, and if bad outcomes never occurred there would be no premium — which means any stretch of time can deliver the risk without the reward.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.