Stocks for the Long Run Ch. 6: Where the Argument Stops — Survivorship, Sequencing, and the Honest Case Against Siegel

Investment Background

The previous five chapters presented Siegel's framework. This chapter states the case against it at full strength.

The reason is the same as in the final chapters of Winning the Loser's Game and When Genius Failed: an overextended conclusion gets discarded along with its correct portion the moment it meets a counterexample. Knowing where the boundary sits is what makes the interior safe to use.

And for Siegel this chapter is especially necessary, because his thesis is the most academically contested of the five books in this phase.

The Wall Street Translation

Criticism One: Survivorship Bias — The Most Serious

This is the strongest and hardest-to-rebut criticism of Siegel.

Siegel's data is American, from 1802 to the present.

Now ask: if you were standing in the world of 1900, choosing one country to invest in for a hundred years, would you have picked the United States?

In 1900 the most powerful economy was Britain. Germany was industrializing rapidly. Argentina was among the wealthiest countries on earth, with income per capita above most of Europe. Russia had a substantial stock market.

A rational investor in 1900 would quite likely not have made America their first choice.

Here is what the following century produced:

Market	Twentieth-century outcome
United States	Became the largest economy; world-leading equity returns
Britain	Empire dissolved; returns markedly below the US
Germany	Two defeats, one hyperinflation; equity market wiped out twice
Russia	After 1917, investors lost 100%
China	After 1949, investors lost 100%
Argentina	Fell from the wealthiest ranks; chronic high inflation
Japan	Peaked in 1989; negative real returns for the following thirty years

The Japan row deserves particular attention, because it is a modern developed-market case. An investor who bought at the Nikkei's 1989 peak had not recovered real purchasing power even after holding more than thirty years.

That directly violates the regularity implied by the Chapter 4 table.

The criticism at its core: Siegel studied the twentieth century's most successful market. Using a winner's history to infer that "equities must revert over the long run" is methodologically like studying lottery winners to infer the expected value of a lottery ticket.

The Dimson, Marsh and Staunton international long-run dataset is the most systematic response to this problem. Their finding: global equities' long-run real return is indeed positive, but clearly below the US figure — roughly 5% — with enormous variation across countries.

Criticism Two: Sample Size

Chapter 4 raised this; it deserves formalizing here.

Two hundred years contains only ten non-overlapping twenty-year periods and six or seven non-overlapping thirty-year periods.

Supporting the claim "thirty years has never been negative" with seven observations is statistically very weak.

And those periods are not independent of one another — they share a country, an institutional system, and a technological history. Statistically this is closer to one observation than to seven.

Criticism Three: Sequence Risk — The Most Lethal for Retirees

This is the criticism that matters most to readers of this site.

Siegel's framework rests on buy and hold to the end of the period. The annualized returns in his tables assume you invest a sum at the start, touch nothing, and withdraw at the end.

Retirees do not do that. Retirees withdraw every year.

And once you are withdrawing, the order of returns begins to matter enormously — something an annualized average completely conceals.

Consider two retirees whose thirty-year average returns are identical:

• A: the market falls sharply in the first five years, then rises for a long time.
• B: it rises for the first five years, then falls at some later point.

In Siegel's framework their thirty-year annualized returns are the same.

In reality A may be broke by year eighteen while B is fine.

Because A was forced to sell shares at market lows to fund living expenses, and those shares can never participate in the later recovery.

This is sequence-of-returns risk, worked through numerically in Chapter 2 of Retirement Decumulation Mechanics — two identical return sets in opposite order diverging by roughly $70,000 within four years.

This must be stated with total clarity: Siegel's long-run data is a valid planning tool for an investor in the accumulation phase. For a retiree in the withdrawal phase it is insufficient, and dangerous used alone.

Criticism Four: "The Long Run" May Outlast You

A simple point, often dodged.

If mean reversion takes twenty years and you retire at 65, the moment this framework pays off may arrive when you are 85.

And you may need the money before then, or not be here.

The statement "equities revert over the long run" is, for an individual, bounded by their actual lifespan and cash flow needs — not by the availability of historical data.

So What Remains

Accepting all four criticisms — and they are all forceful — what remains of the framework is clearer and more trustworthy:

1. Equities represent ownership of real productive assets, so their cash flows reprice with inflation. This is a mechanical conclusion that does not depend on US historical data. It holds in any country.

2. Nominally safe assets face real purchasing-power risk over long horizons. The 1940–1981 bond record establishes this, and the mechanism (fixed nominal cash flows meeting inflation) is universal.

3. Return dispersion narrows with holding period, because the valuation term is diluted while dividends and earnings growth are relatively stable. That mechanism is robust even if the specific American table is optimistic.

4. Starting valuation shifts the distribution of the following decade's returns. This holds in international data too — in fact more clearly than in US data.

5. Plan in real returns rather than nominal ones. This depends on no contested premise whatsoever.

Those five survive even if Siegel's central thesis is substantially weakened. They are this book's genuinely safe output.

Executable Trading Rules

1. Weight international diversification more heavily than Siegel does. This is the most direct practical corollary of the survivorship criticism. If you cannot be certain your country is the next America, hold the globe. A globally diversified equity fund is the direct implementation of honestly admitting you do not know which country wins.

2. Use Siegel's numbers in accumulation; never alone in withdrawal. This is the chapter's most important rule. Accumulation: planning savings at 6–7% real is reasonable. Withdrawal: you must layer sequence risk analysis on top, using Monte Carlo or historical path simulation rather than an annualized average. Our Advanced Withdrawal Simulator and Retirement Cash Flow Analyzer exist for this.

3. Test the plan against the worst case, not the average. A recurring rule, but this chapter adds a new reason: the average may itself be inflated by survivorship bias.

4. Lower the return assumption you take from this book, as a survivorship correction. A practical approach: use 5% rather than 6.5% as your real return assumption for global equities. That corresponds roughly to the Dimson international data rather than the American special case.

5. Do not abandon equities because of this chapter. This is the most important boundary. Every criticism above weakens the strong claim that "equities must revert over the long run." None of them weakens — or can weaken — the mechanical conclusion that over multi-decade horizons, only ownership of real productive assets defends against the erosion of purchasing power. The correct use of these criticisms is to lower expectations and broaden diversification, not to retreat to cash.

Relevance to a Retirement Portfolio: Closing

Six chapters, five sentences:

• Chapter 1: equities returned about 6.5% real over the long run, with a stability that spans two centuries of extreme institutional change.
• Chapter 2: bonds are safe over short horizons and unsafe over long ones, because "safe" denotes two different properties at one year and at thirty.
• Chapter 3: the bulk of that return comes from reinvested dividends, not price appreciation.
• Chapters 4–5: dispersion narrows as the horizon lengthens, but your purchase valuation shifts the whole distribution.
• Chapter 6: and all of it rests on the history of the most successful country, and does not fully apply to a retiree who must withdraw.

This book's place in the library is specific: A Random Walk Down Wall Street tells you how to hold equities (low-cost index funds), Winning the Loser's Game tells you why you cannot win by selecting (market structure), and this book tells you what equities have actually delivered over the long run, and how far that number should be trusted.

Its practical output is a number for planning, not a trading strategy.

Our standard position is unchanged: low-cost, globally diversified index funds as the core, with a cash buffer covering essential spending, planned in real returns and tested against the worst case.

Siegel's data supports that conclusion. So do this chapter's criticisms — they simply require you to set expectations lower and diversify wider.