Stocks for the Long Run Ch. 3: Where the Return Actually Comes From — Dividends and the Reinvestment Engine

Investment Background

Chapter 1 said the real return on equities is about 6.5%. This chapter takes that number apart.

Once it is apart, you will find that most people have the source of equity returns backwards.

When people think about "making money in stocks," they picture price appreciation — buy at 100, it rises to 150.

But over the long run, price appreciation is not the main source of return.

The Wall Street Translation

Three Components

Long-run total equity return decomposes precisely into three pieces:

One, the dividend yield. Companies pay out part of their profits in cash. Historically, the average US dividend yield ran roughly 4–5% (markedly lower in recent decades, which we address below).

Two, real earnings growth. Company profits grow with the economy. Historically about 1.5–2% real per year.

Three, the change in valuation. The multiple the market will pay per dollar of earnings shifts. Over very long horizons this term approaches zero — because multiples cannot expand or contract indefinitely.

Add the first two: roughly 4.5% plus 2%, which is about 6.5%.

That is where the Chapter 1 number comes from. It is not mysterious. It is dividends plus earnings growth.

Why the Decomposition Matters

Because it tells you that over the long run valuation change is irrelevant, and over the short run valuation change is nearly the only thing that matters.

Within a single year, a 30% move is almost entirely a change in the multiple — people's mood. Earnings do not change 30% in a year.

Over thirty years, the multiple change is diluted toward zero, and what remains is dividends and earnings growth.

That insight has a direct practical corollary:

The longer you hold, the more your return depends on actual business operations and the less it depends on other people's emotions.

This is the structural reason long-term investing beats short-term trading, and it requires no assumption about market efficiency at all.

The Compounding Effect of Reinvestment

Now for the chapter's most counterintuitive part.

The dividend alone is only 4–5%. That sounds modest. But when reinvested, it buys more shares, and those shares produce dividends of their own.

Over thirty and fifty years, that loop produces an enormous effect.

Siegel's well-known comparison sets the price index against the total return index — the curve without dividends against the curve with dividends reinvested. Over a century, the gap between them is an order of magnitude.

Concretely, assuming 2% real earnings growth and a 4.5% dividend yield:

Holding period	Price growth only (real)	Price plus reinvested dividends (real)
10 years	about 1.22x	about 1.88x
30 years	about 1.81x	about 6.61x
50 years	about 2.69x	about 23.3x

Over fifty years, reinvested dividends produce more than eight times the terminal wealth of price appreciation alone.

The majority of long-run equity return does not come from rising share prices. It comes from dividends, and from the dividends paid by the shares those dividends bought.

A Modern Problem That Must Be Handled Honestly

The historical figures above carry an important modern qualification, and omitting it would be dishonest.

In recent decades the US dividend yield has been well below its historical average, typically 1.5–2%.

Does this invalidate Siegel's framework?

No, but it requires an adjustment.

The reason is that how companies distribute cash has changed. A large share of the cash that would once have been paid as dividends is now returned through share buybacks.

A buyback is economically similar to a dividend — the company uses cash to reduce share count, earnings per share rise, and remaining shareholders own a larger proportion.

So the correct measure is total shareholder yield: dividend yield plus net buyback yield. On that basis, the modern figure is much closer to the historical one.

But two things must be conceded:

1. Buybacks are not fully equivalent to dividends. Their timing is set by management, and managements tend to buy back when the share price is high — which disadvantages the continuing shareholder. Dividends are mechanical.
2. Even measured as total shareholder yield, the contemporary level may still run slightly below history. This connects to the valuation question in Chapter 5.

Division of Labor With the Rest of the Library

This chapter is easy to confuse with several existing books, so the boundaries need stating:

Book	What it says about dividends
A Random Walk Down Wall Street	Capture market return via low-cost index funds (the cost argument)
Value-investing titles	Use high dividend yield as a valuation screen (a stock-selection argument)
This book	Reinvested dividends are mathematically the bulk of long-run total return (an accounting identity, not a selection recommendation)

This must be explicit: this chapter is not advising you to buy high-dividend stocks.

High-dividend strategies are an active selection decision carrying their own risk — high yields frequently appear in structurally declining industries.

What this chapter states is an accounting fact at the whole-market level: most of total return comes from reinvested distributions. Its practical corollary is about reinvestment, not about selection.

Executable Trading Rules

1. Make sure your dividends are automatically reinvested. This is the single most important and most easily executed line in the chapter. At most brokers it is a checkbox. Outside taxable accounts, turn it on. Those seemingly trivial quarterly distributions are the bulk of your wealth in thirty years.

2. In taxable accounts, note the tax consequence of reinvesting. Dividends are a taxable event on receipt even if immediately reinvested. This is why high-yielding assets usually belong in tax-advantaged accounts — the account location rules are in Chapter 5 of Retirement Decumulation Mechanics.

3. Always look at total return, never the price index. When media report "the market went nowhere for a decade," they are usually quoting the price index. With dividends included, that same decade is often positive. This distinction matters most when sentiment is worst.

4. Understand that in retirement your dividends are cash flow. This planning tool is routinely overlooked. A retirement portfolio in total-market index funds itself throws off 1.5–2% in cash distributions. That portion of your withdrawal requires selling no shares at all, and so is insulated from sequence-of-returns risk.

5. Do not let this chapter send you chasing high-dividend stocks. See the division of labor above. A market-level accounting fact is not a stock-level selection recommendation.

Relevance to a Retirement Portfolio

This chapter supplies the mechanical explanation behind our core recommendation across this site.

Why do we consistently recommend low-cost total-market index funds? Not only because they are cheap (Malkiel's argument), and not only because you cannot outplay the professionals (Ellis's argument).

Also because they plug you directly into the engine this chapter describes: you own every dividend the whole market pays, they reinvest automatically, and you need exercise no judgment.

For those already retired, this chapter supplies a concrete planning element:

If your portfolio yields 2% and your withdrawal rate is 4%, half your withdrawal comes from cash flow and only the other half requires selling shares. That directly reduces the pressure to sell in a down year.

This does not replace the cash buffer — Chapter 2 and Chapter 3 of Retirement Decumulation Mechanics establish that the buffer remains necessary. But it reduces how much work the buffer has to do.

Chapter 4 returns to the question Chapter 1 left open: 6.5% is a long-run average, so how bad can the short run be, and how long does "long run" actually have to be?