Expected Returns Ch. 6: Where the Framework Stops

Investment Background

The first five chapters built a framework. This chapter marks where it stops.

Especially necessary for this book, because its register is quantitative and precise — and precision itself manufactures unwarranted confidence.

Against the Gods Chapter 6 established the warning: the market pays for precision, not for honesty. This chapter applies that standard to Ilmanen himself.

The Wall Street Translation

Limit One: Premia Are Estimates, Not Entitlements

The most important one.

The phrase "equity premium" invites the belief that it is a promise — bear the risk and the compensation arrives.

It is not a promise. It is an average estimated from historical data, and estimates carry error.

Bernoulli's law from Chapter 3 of Against the Gods applies directly: improving an estimate's precision tenfold requires a hundred times the data. And the error band around the equity premium estimate is uncomfortably wide — academics still do not agree on what the true equity premium is.

More fundamentally, Knight's distinction from Chapter 5 of Against the Gods: a premium's existence assumes the future institutional environment resembles the one that produced the historical data. That assumption is itself uncertainty, not risk.

The Japan case in Chapter 6 of Stocks for the Long Run is the sharpest reminder: for more than thirty years after 1989, Japanese equity investors bore the full equity risk and the premium did not arrive.

Limit Two: An Observed Premium Can Be Crowded Away

This one is especially damaging to the book's framework.

If a premium becomes widely recognized, capital flows in, prices rise, and the future premium falls.

This is not theory — it is the definition of arbitrage, and Chapter 3 of When Genius Failed describes the process: edges decay under competition.

So there is a self-negating property: the more successful this book and the more people use the framework, the smaller the premia it describes.

A concrete example: several factor premia documented in academic research have performed materially worse after publication than in their pre-publication backtests. The phenomenon has itself been studied and documented.

This does not mean the foundational premia (equity, term) vanish — they correspond to real discomforts that cannot be arbitraged away. But it means the more elaborate, identification-dependent premia deserve a discount.

Limit Three: Half This Book Does Not Apply to You

Chapter 1 previewed this; it needs full statement here.

Much of Ilmanen's book concerns constructing a portfolio in which each premium contributes equally, and achieving that usually requires leverage.

The risk parity logic: bonds are less volatile than equities, so equalizing their risk contribution requires levering the bonds.

That logic is mathematically sound. And it does not apply to a retirement investor, for reasons this library documents in full:

When Genius Failed devotes a whole book to leverage's real cost — not magnified losses but the elimination of the option to do nothing. And for a retiree, that option is the most valuable thing they own.

So the correct reading is: take the diagnosis, leave the implementation.

Limit Four: Precise Language Conceals Crude Estimates

This one concerns presentation.

When someone says "the term premium is currently 0.8 percentage points," that decimal conceals a fact: it was estimated from a model with an enormous error band.

Where Are the Customers' Yachts? Chapter 2 explains why: precision outsells honesty.

So the correct posture when using this framework is: accept the direction, distrust the decimal.

"Expected equity returns are currently below the historical average" is a useful, robust statement. "Expected real equity returns are currently 4.2%" is false precision.

So What Remains

Accepting all four limits, what remains is real and extremely useful to a retirement investor:

1. Returns are a stack of premia, each layer corresponding to a nameable discomfort. (Ch. 1, 3) This structural insight depends on no specific numerical estimate.

2. The risk-free rate sets the floor for everything, and the real rate determines your actual position. (Ch. 2) Observable, no model required.

3. Set planning assumptions from current yields rather than historical averages. (Ch. 1, 2) Nearly certain for bonds; for equities at least closer to your actual starting point than an average.

4. Measure diversification by premia rather than by number of products. (Ch. 4) Purely diagnostic, requiring no forecast.

5. Use this information for planning rather than trading. (Ch. 5) And mechanical rebalancing captures part of it for free.

Those five hold even if every specific premium estimate is wrong. They are the book's genuinely safe output.

Executable Trading Rules

1. Accept the direction, distrust the decimal. The chapter's most practical line. "Below the historical average" is usable; "4.2%" is false precision.

2. Treat foundational and elaborate premia differently. The equity and term premia correspond to real, unarbitrageable discomforts and are relatively reliable. Factor premia requiring complex identification deserve a discount (Limit Two).

3. Never use leverage to "equalize" your premium contributions. Absolute. The risk parity mathematics is correct, and it does not apply to someone who cannot survive being forced to sell.

4. Update your planning assumptions annually, not quarterly. This information changes slowly (Chapter 5). Looking more often only tempts you to treat a planning tool as a trading signal.

5. Treat the framework as diagnostic spectacles, not a forecasting machine. It lets you see what your portfolio is made of — a use independent of the accuracy of any numerical estimate.

Relevance to a Retirement Portfolio: Closing

Six chapters, four sentences:

• Chapters 1–2: your return is not a number the market hands you but a stack of risk premia whose floor is set by the risk-free rate — and every layer is observable today, which sits closer to your actual starting point than a historical average.
• Chapter 3: every premium compensates a specific discomfort, and nearly all of them become painful in the same bad times.
• Chapters 4–5: so diversification should be counted in premia rather than products — and this information belongs in planning rather than trading, with mechanical rebalancing capturing part of it for free.
• Chapter 6: but premia are error-laden estimates rather than entitlements, and the institutional half of this book does not apply to you.

This book's place in the library is specific:

Stocks for the Long Run tells you what equities delivered historically. Mastering the Market Cycle tells you which end of the pendulum you occupy. Against the Gods tells you what can be measured and what cannot.

This book tells you what each part of your portfolio is being paid to bear, today.

Its practical output is the number you type into a retirement calculator, plus an honest inventory of how many distinct risks you actually hold.

And our standard position gains a more precise formulation after this book:

A low-cost, globally diversified index fund as the core — bearing the equity premium, the most reliably compensated one and the one required against inflation.

High-quality intermediate bonds — a moderate term premium, usually moving opposite equities in a growth shock.

A cash buffer covering one to three years of essential spending — deliberately forgoing the liquidity premium in exchange for having something to sell when everything else has fallen.

Plus a portion in inflation-linked bonds — for the scenario in which bonds cannot protect you.

Four holdings, four nameable premia, painful moments spread across different economic scenarios, no leverage anywhere, and no forecasting required.

Ilmanen argued the structure of premia across a heavy institutional textbook. For you, that structure's conclusion is simple enough to write on one page — and it agrees with every other book on this site.