A dollar put into large-capitalization US stocks in 1926 was worth $14,751 a hundred years later, a compound return of 10.1% a year.
The yearbooks that carried that record no longer exist. Morningstar stopped producing the stocks, bonds, bills and inflation data behind them, so Roger Ibbotson rebuilt the series from scratch for its hundredth year, and added the centuries before 1926 and the markets outside the United States.
"You don't have 14,000 times the purchasing power today that you would have had in 1926 by buying and holding the US market."
Ibbotson wrote the 1976 articles that first measured 50 years of US stock and bond returns, built the data with Rex Sinquefield while he was at the University of Chicago, and is the main author of the book this conversation is about.
I listened to the full interview so you can skip it. 34 minutes of audio, 15 minutes of reading.
Here are the 12 takeaways that matter.
👤 Guest: Roger Ibbotson, professor emeritus of finance at Yale School of Management and chairman of Zebra Capital Management, who built the original stocks, bonds, bills and inflation return series in 1976
🎙️ Host: Lotta Moberg, an economist who is cofounder and head of investment research and strategy at Wealth Horizons
👥 Also on: Laurence Siegel, the Gary P. Brinson director of research, emeritus at the CFA Institute Research Foundation and the executive editor of the book
📰 Published: 1 September 2026 on the Financial Thought Exchange podcast feed, and 27 August 2026 on YouTube (CFA Institute Research Foundation)
🔴 YouTube | 🟢 Spotify | ⏱️ 34 min | ✅ Time saved: 19 min
Key Takeaways
A dollar in US large-capitalization stocks became $14,751 over 100 years, a return of 10.1% a year Inflation of 2.9% a year cuts that to about 800 times the purchasing power Ibbotson says the real return works out near 7%, or a doubling every 10 years
The data set behind the number had been discontinued and had to be rebuilt Morningstar stopped producing the stocks, bonds, bills and inflation data The rebuilt version adds the centuries before 1926 and markets outside the United States
The US was the best-performing market in the world over the century, which makes it the wrong one to extrapolate Siegel says a global investor would still have done well, because some countries do badly in any given period
None of this return required skill, and almost none of it was capturable before the 1970s Index funds only arrived around 1976, and reinvesting dividends was hard before that
Dividends have fallen from 3.7% of price to a little over 1%, replaced by buybacks Halving the compounding rate does far more than halve the ending value
Taxes, not fees, are now the dominant variable for a taxable retail investor Ibbotson names two legal escapes from capital gains tax: never sell, or die
Not every investor can stay fully invested, so most earn something closer to the economy's growth Money taken out of the market moves you toward 3% real rather than 7%
Buybacks are shrinking share counts while a large new-issue market adds to them Ibbotson says the market is not running out of shares; prices, not share counts, carry the wealth
Stock returns and GDP growth do match in the long run, whatever the year-by-year studies say Equities are a levered version of the economy, so they outperform it slightly
1. The SBBI data was rebuilt
The book exists because the data behind it stopped being published. Ibbotson said the trigger was the 50th anniversary of his original 1976 articles, the first of which measured historical returns from 1926 for 50 years. The second made forecasts out to the year 2000, and those have now been extended to 2025 so the projections can be scored against what happened.
The stocks, bonds, bills and inflation yearbooks were discontinued and Morningstar stopped producing the underlying data, which is what prompted a rebuild rather than an update. Every index in the set has been reconstructed, giving a full century of US stock and bond returns
The project also brought in co-authors to cover the markets before 1926 and the non-US markets, so the book runs from a century of US data out to several centuries of global data
Ibbotson said nobody could see the shape of this before the original work. "They had no idea of exponential wealth actually." Earlier work existed — he pointed to the studies by Fisher and Lorie — but it had not been updated, and the comparison between stocks and bonds was not established
The title is the claim: money held over long periods grows exponentially rather than by addition
2. The US was the best market
Moberg asked Siegel, who edited the volume and worked with the contributing authors, what surprised him most.
Siegel's answer was two findings that pull against each other: how similar markets have been across centuries and countries, and how different. The mechanics of investing for future growth have not changed at all, even as market institutions have
The US is the outlier, and treating it as the base case is the error. "For example, the US was the best performing market in the world over the last century." "If you were to extrapolate US results to the whole world in the future, you'd get an awfully high estimate and you'd probably be over optimistic."
He added that the world market still grew at a respectable rate, enough to make a diversified long-term investor wealthy, because a country that does badly in one period is offset by others
3. A dollar became $14,751
Ibbotson walked through the chart of a dollar's growth over the past 100 years, which the show displayed for viewers.
The headline number is the compounding, not the return. "a dollar invested in just in large cap stocks grew to 14,751" — almost 15,000 times the starting dollar, from a return he put at 10.1% a year
The rule of 72 is how he gets from one to the other: at about 10%, the market doubles roughly every seven years, so one becomes two, then four, then eight Moberg took the audience through the arithmetic herself. "You take 72 divided by whatever that growth number is."
The chart is plotted on a logarithmic scale, with the stock market, the bond market and inflation on it. Ibbotson explained why that matters: on a log scale the slope of the line is the return, so a reader can pick any stretch of the century and read the return off the graph
The series is the US market. International returns were a bit lower, he said, but still high
4. Inflation leaves 800 times
The nominal figure is not what an investor could spend. "It's about 800 times because inflation was about 20 to 1", which brings the real value of the number down
Inflation over the period ran at 2.9% a year
Ibbotson did not treat the smaller number as a disappointment. "But 800 times your money is a huge amount of money and it does not in any way change the story that if you're a long-term investor that you should do very well."
Moberg put the real return at roughly 7% a year and asked whether that was fair; Ibbotson said it was roughly correct, which on the same rule of thumb is a doubling of purchasing power about every 10 years
The book's forecasts run to 2050 and are distributions, not point estimates. They are built by bootstrapping and Monte Carlo analysis — repeatedly redrawing from the historical record to produce a range of outcomes rather than one number — and the final chapter uses international risk premiums rather than US ones to avoid the extrapolation problem Siegel described
5. Index funds arrived in 1976
Moberg asked what a century of index returns means for a person who has to hold the index for that long.
Ibbotson was explicit that the return carries no skill component. "There's no extra return. This is just being in the stock market, being in large cap stocks."
For the first half of the sample it was not practically capturable. Holding a diversified portfolio of large-capitalization stocks and reinvesting the dividends was difficult before the 1970s
The instrument arrived at almost exactly the moment the research did. Vanguard's index funds were coming out around 1976, Ibbotson was at the University of Chicago and had worked on some earlier indexes, and he said he bought the funds himself
On cost, he would not guess at the 1970s number and then gave the range anyway. "Well, right now it's like less than 10 basis points mostly, but then it might have been 25 or something." Index funds were always low priced, he said
Early funds did not hold every stock. Ibbotson said they used stratified sampling to build something that closely resembled the S&P 500; only later did buying all the constituents become cheap and easy
6. Holding through the 1930s
Moberg asked how a well-connected investor in the 1920s, 1930s and 1940s would have stayed in the market in practice.
Buying and holding was possible; reinvesting the dividends was the hard part. Plenty of people did hold their positions
The start date was not chosen for narrative reasons. The University of Chicago's Center for Research in Security Prices databases began in 1926, so Ibbotson's studies did too — work he did originally with Rex Sinquefield — and that start date pulled the Depression into the record
Many of the investors who came to the market in the 1920s were badly hurt in the 1930s, and it took time for retail buyers to return
The people who did not sell are the whole point of the series. "Although it wiped many people out because they were leveraged or they were selling in order to raise money to live on, if you were able to hold the market through that period, you did extremely well." Those who stayed invested were there for the bull markets after the Second World War
7. Total returns measured late
The returns are tied to the economy through what companies supply and what investors demand, which is the framework Ibbotson said he has worked on for years: demand covers what people will pay and how much risk and illiquidity they will tolerate, supply covers what companies in the economy produce
Real GDP growth of around 3% is lower than the market's return because most of national output is consumed rather than reinvested, while the return series assumes every dividend goes back in
Measuring the two halves together is recent. Until the late 1960s, capital gains and income were reported separately and not added up. "And hardly anything had been measured in total return form prior to these studies and even the index funds didn't start out having reinvestment programs."
The index measures what the market offered, not what any individual achieved. Different investors withdraw different amounts, and an index that tried to allow for withdrawals would be describing one investor rather than the market
Payouts have shifted from dividends to buybacks, which Ibbotson attributed mainly to taxes. Either way the cash leaves the company, and an investor who wants cash can sell shares
8. Dividends fell to about 1%
The share of the return that is paid out has collapsed. "It was historically 3.7% of the yield on the large cap stocks. More recently it's just a little over 1% because as I say they're replacing this with buybacks."
Taking money out does not cost you proportionally, it costs you exponentially. Ibbotson's example is the doubling time: going from doubling every seven years to doubling every 14 years produces a far lower ending value "Not going from 10% to 5% doesn't decrease your value at the end by half. It dramatically changes the value."
Moberg stopped to define the term for the audience, separating the dividend yield from the payout ratio: the yield is the dividend measured against the share price, not the share of earnings a company distributes, which is why 3.7% sits so far below the other numbers in the conversation
She also put the mirror-image question to Ibbotson: if companies pay out less, the economy has less reinvested capital, but investors receiving dividends can put them back into companies themselves
9. Taxes decide what you keep
Tax rates used to be much higher, Ibbotson said, but there were far more loopholes then and there are still plenty now, which is how large fortunes end up paying effective rates no higher than a retail investor's
He named two ways to avoid capital gains tax and treated one of them as a joke. "And so but one simple way to get around capital gains tax is to die because there's a step up on your death." The other is never to sell, because an unsold stock is not taxed
Siegel raised the fund wrapper, and Ibbotson said exchange-traded funds have better tax characteristics than mutual funds — one of the main reasons money has moved from mutual funds into ETFs
For a taxable investor, tax management now outweighs the other costs. Fees and implementation costs exist, but Ibbotson said how much you reinvest and how you handle taxes is what determines whether you actually get the compounding in the book
Tax-deferred accounts are the exception that makes the historical return reachable: the money compounds at the rates in the book and the tax is paid once, on withdrawal
10. Not everyone can stay in
The 10% return assumes full reinvestment, and everybody cannot do that at once. Whether companies pay dividends or buy back stock, the cash leaves the corporate sector and someone ends up holding it
In aggregate, investors consume what companies produce; the index measures the production, not the consumption
The gap between the two numbers is the withdrawal. "And so to the extent we take money out of the market, we more become more like the overall economy" — closer to about 3% real growth than to the 7% real return on stocks
Ibbotson's framing is that the total return is what corporations create, and that each investor then decides how much of it to take out
11. Shares are not running out
Moberg asked whether a shrinking pool of shares changes the picture for someone who wants to own more of the economy.
Ibbotson said supply is not the constraint. "Well, it's not like we're running out of shares." Additional demand shifts the demand curve outward and raises prices, which is a price signal that what these companies produce has become more valuable
Buybacks and new issues run at the same time. Buybacks reduce the share count; meanwhile there is a large initial public offering market, and he named SpaceX as an example of the new issuance going on now. Whether the total share count is rising or falling is not clear and varies by year
The wealth can grow without the share count growing, because one share worth $1 and one share worth $100 are not the same claim
For an index builder, he said, the answer to all of it is to measure the total return, since each investor takes out a different amount of cash
12. GDP and returns do match
Moberg closed by putting the standard objection to both of them: economists have published work finding no relationship between GDP growth and stock returns.
Ibbotson said the connection is there and has been fairly tight, once you account for the fact that most of GDP is consumed while most of the stock market's return is reinvested. His own studies find the two match in the long run after that adjustment
Equities move more than the economy because they are geared to it. "So as a more levered version it outperforms a bit"
The reason the studies come out the way they do is measurement noise, not the absence of a relationship. "One reason why they might not look like they match each other all that closely is that the stock market is very noisy." Share prices move all over the place while GDP growth does not, so the timing of market gains does not line up with the timing of growth quarter by quarter
His conclusion is that a paper can be right and wrong at once: "you could write a paper saying that they're not related but they're very closely related in the long run"
Bonus Insights
Siegel's own holding is the episode's best illustration of the arithmetic. He bought the Vanguard 500 index fund in 1983 with an insurance payout after he crashed his car, still owns the shares, and put them at "a hundred times what I paid for them"
Moberg noted she thinks about this material as an economist, and pressed both guests repeatedly on what a century of returns says about economic growth rather than about portfolios
This is the first of two conversations. The second covers index construction and the methodology behind the rebuilt series, which is where the numbers in this one come from
The book itself is a compendium rather than a monograph: Ibbotson is the main author, Siegel the executive editor, and a long list of finance academics contributed chapters on individual markets and periods
Ibbotson's bottom line is that the century-long record is real and reachable, but only for an investor who reinvests everything, holds through the 1930s-style falls, and manages the tax bill — and that the closer you get to spending the income, the closer your return gets to the economy's 3% rather than the market's 10%.
Products, Companies & Tools Mentioned
Morningstar (Stopped producing the stocks, bonds, bills and inflation data, which is what prompted the rebuild behind the book)
Zebra Capital Management (The firm Ibbotson chairs, alongside his emeritus post at Yale School of Management)
CFA Institute Research Foundation (Publisher of the book and of this podcast; Siegel is its director of research, emeritus)
Vanguard and the 500 Index Fund (The index funds Ibbotson says were arriving around 1976 and bought himself, and the fund Siegel has held since 1983)
Center for Research in Security Prices, University of Chicago (The database whose 1926 start date set the start date of the return series, and with it the inclusion of the Depression)
S&P 500 (What early index funds tracked by stratified sampling rather than by buying all 500 constituents)
SpaceX (Ibbotson's example of the large new-issue market running alongside corporate buybacks)
Books & Resources Mentioned
Exponential Wealth: Centuries of Stock and Bond Returns – Roger Ibbotson, with Laurence Siegel as executive editor (The publication the conversation is built on: a century of rebuilt US data, the centuries before it, the global markets, and forecasts to 2050)
Stocks, Bonds, Bills, and Inflation (SBBI): 2020 Summary Edition (The yearbook series Ibbotson says has been discontinued; from the episode notes)
Stocks, Bonds, Bills, and Inflation: The Past and the Future, 1982 Edition (The earlier edition carrying the 50-year forecasts he says have now been scored against 2025; from the episode notes)
Stocks, Bonds, Bills, and Inflation: Historical Returns (1926–1987) (The historical-returns edition of the same series; from the episode notes)
Rates of Return on Investments in Common Stocks – Lawrence Fisher and James Lorie (The earlier University of Chicago work Ibbotson says existed before 1976 but had not been updated)
The Long-Run Drivers of Stock Returns: Total Payouts and the Real Economy (The paper the episode notes point to on the payouts-and-growth question that closes the conversation)
The Active Side of Indexing and Index Construction and Fundamental Indexing, both with Rob Arnott (Earlier conversations on this show that the notes point to, covering the index-construction ground part two takes up)
If this was worth your time, send it to someone who has to have a view on this.
Get the latest market chatter as it happens: