$10,000 put into US shares in January 2006 became $88,249 by October 2026. The same $10,000 in long-term US government bonds became $17,368. Both figures include dividends, the cash that shares and bonds pay out, but not costs or tax.
Shares paid more. Bonds still beat them in five of the twenty calendar years (January to December) from 2006 to 2025. In 2008 the bonds gained 33.9%. Shares lost 36.8%. In 2022 both fell, but the bonds fell further.
Do stocks or bonds pay more? US shares paid more. SPY (the S&P 500 ETF, a fund that trades like a share) earned 11.03% a year from 3 January 2006 to 26 October 2026. TLT (the long-term US Treasury bond ETF) earned 2.69%. Both include dividends. The extra return shares earn over bonds is the equity risk premium.
Shares earned 8.34 percentage points a year more than the bonds. Their biggest fall from a previous high was 55.2%, against 48.3% for the bonds.
Bonds beat shares in five of the twenty calendar years, including 2008. Both fell together in 2018 and 2022.
The bond fund has not regained its August 2020 high. It sits 42.1% below it.
What are we comparing?
Shares (also called stocks) are part-ownership of companies. Their payout is not fixed. A bond is a loan that pays fixed interest and returns the loan at the end. Our explainer on bond prices and yields covers how those prices move.
For shares we use SPY, an exchange-traded fund (ETF) that follows the S&P 500, the index of large US companies.
For bonds we use TLT, an ETF that follows US Treasury bonds with more than twenty years left to run, according to the fund issuer's own page. Our explainer on what an ETF is covers the structure.
So "bonds" in this post means long-term US Treasury bonds.
Which paid more, stocks or bonds?
Source: YX Insights price data
The chart shows what $10,000 became in each fund. Shares end at $88,249, the bonds at $17,368.
That is 8.82 times the money for shares, or 11.03% a year. The bonds made 1.74 times the money, or 2.69% a year.
Those yearly rates are compound annual rates. A compound annual rate is the single yearly return that turns the starting sum into the ending sum. Our explainer on why the compound rate and the average differ covers it.
The gap, 8.34 percentage points a year, is the equity risk premium measured after the fact in this window. The next section explains the term.
The equity risk premium is the extra return that shares earn over bonds. "Equity" is another word for shares.
Shares have no fixed payout, while a bond promises fixed interest. The premium is the extra return that went with holding that uncertainty. Our explainer on the equity risk premium measures it over a much longer history.
How often did bonds beat stocks?
Source: YX Insights price data
Each pair of bars is one calendar year. The chart includes 2026 as a part year to 26 October, while 2006 starts at the close of 3 January. The count of twenty uses 2006 to 2025.
Bonds beat shares in five of those twenty years: 2007, 2008, 2011, 2014 and 2018. Shares won the other fifteen.
The two biggest bond years were 2011 and 2008. In 2011 the bonds gained 34.0%, while shares gained 1.9%. In 2008 the bonds gained 33.9%, while shares lost 36.8%.
The pattern also ran the other way. In 2009 shares gained 26.4% and the bonds lost 21.8%.
The bonds lost money in seven of the twenty years, against three for shares. In 2018 the bonds "won" only because they fell 1.6% while shares fell 4.6%.
So far in 2026, to 26 October, shares have gained 13.1% and the bonds have lost 3.0%.
How much did prices swing and fall?
Volatility is how widely a price swings from day to day, scaled to a year. SPY's was 19.17% and TLT's was 14.73%.
A fall here is measured from a previous high to the lowest point after it. We searched the whole window, 3 January 2006 to 26 October 2026, on daily closes.
Source: YX Insights price data
The chart plots how far each fund sat below its previous high, week by week.
Shares fell 55.2% between 9 October 2007 and 9 March 2009. That is about 17 months. They regained the old high on 16 August 2012, 4.85 years after it.
The bonds fell 48.3% between 4 August 2020 and 19 October 2023. That took 3.21 years. They have not regained that high. On 26 October 2026 they sit 42.1% below it.
The bond fall was slow. The fund has now spent 6.23 years below its high, against 4.85 years for shares.
US shares (SPY) | Long-term US Treasury bonds (TLT) | |
|---|---|---|
Value of $10,000 on 26 October 2026 | $88,249 | $17,368 |
Times the money | 8.82 | 1.74 |
Annual rate | 11.03% | 2.69% |
Volatility, per year | 19.17% | 14.73% |
Worst fall from a high | −55.2% | −48.3% |
The high before it | 9 October 2007 | 4 August 2020 |
The low | 9 March 2009 | 19 October 2023 |
Old high regained | 16 August 2012 (4.85 years after the high) | Not by 26 October 2026 (6.23 years and counting) |
$10,000 invested at the close on 3 January 2006, dividends counted, before costs and tax. Falls are searched over the whole window, 3 January 2006 to 26 October 2026. Source: YX Insights price data
The table sets the returns, swings and worst falls of the two funds side by side.
What happened in 2022, when both fell?
In 2022 shares lost 18.2% and the bonds lost 31.2%. Our explainer on the 2022 rate hikes gives the month-by-month path.
When interest rates rise, the prices of existing bonds fall. The longer a bond has to run, the more its price falls. Almost all of TLT's holdings have more than twenty years left to run.
2018 was the only other year in the twenty when both fell, but that fall was small. Shares lost 4.6%, while the bonds lost 1.6%.
Correlation measures whether two series move together. -1 means opposite directions, +1 means in step, while 0 means no pattern. We use month-end returns.
From February 2006 to December 2021, the correlation between SPY and TLT was -0.30 over 191 months. From January 2022 to 26 October 2026, it was 0.56 over 58 months. We split at 2022 after seeing the result, so this describes the window and is not a test.
For the two combined in one portfolio, see our explainer on the 60/40 portfolio.
How far does the result depend on the window?
We repeated the comparison from the first trading day of every January from 2003 to 2020, ending on 26 October 2026 each time. Shares had the higher annual rate in all 18.
Across those starts, SPY's annual rate ran from 10.82% to 17.09%. TLT's ran from -4.47% to 3.25%.
Start at the bond fund's own high on 4 August 2020 and $1 in shares became 2.53 by 26 October 2026, while $1 in the bonds became 0.58.
The window is long, but it is one window in one country's market.
How did we run the test?
TLT's price history starts in July 2002. We start on the first trading day of 2006 and end on 26 October 2026. That is 5,236 trading days over 20.8 years. The previous section tests other start dates.
"With dividends counted" means the payouts are treated as reinvested. TLT pays out the interest it earns. The price data records those payments as dividends, so the bonds get the same treatment. Nothing is adjusted for inflation, costs or tax.
The test covers one share fund and one bond fund. It does not include cash, Treasuries with only a few years left to run, corporate bonds or other countries. Our guide to asset classes puts stocks and bonds beside cash, commodities and crypto.
So which paid more?
In this window, shares earned 8.34 percentage points a year more than long-term Treasury bonds. Their biggest fall from a previous high was 55.2%. The bonds swung less from day to day, but their biggest fall was 48.3% and they had not recovered by 26 October 2026.
Get the next explainer
The Academy is free to read, here on the web. The Long Trend, our free weekly letter, comes by email.
Learn more with YX Insights
This explainer is part of the YX Insights Academy. Each one takes a single idea and checks it against real data.
The same approach runs through everything else we publish:
Systematic Portfolio: ready-made portfolios for Macro & Megacaps and for Commodities. We publish the holdings and every change.
Multi-model Signals: a daily long-or-flat call on every name we cover. Each call shows how strongly our four models agree.
Research: company deep dives, macro commentary and essays on how we test.
Good places to start on the website:
Academy: more explainers like this one.
Start here: get our free weekly letter by email.
Macro & Megacaps Systematic Portfolio and Commodities: the two portfolios in detail.
Track record: live results since 3 August 2026, plus the ten-year backtest.
Upgrade: get the Systematic Portfolio, Multi-model Signals and research in full.
Common questions about stocks vs bonds
Do stocks beat bonds over 20 years?
Yes, in our test. US shares (stocks), measured by SPY, grew at 11.03% a year from 3 January 2006 to 26 October 2026, with dividends counted. Long-term US Treasury bonds, measured by TLT, grew at 2.69% a year. Shares ended ahead from all 18 January start dates tested between 2003 and 2020.
The equity risk premium is the extra yearly return that shares (equity) earn over safer-looking government bonds. Shares have no fixed payout, while a bond promises fixed interest. In our window, the gap was 8.34 percentage points a year: 11.03% for shares against 2.69% for long-term Treasury bonds.
Do bonds go up when stocks go down?
Not reliably. In 2008 shares lost 36.8% and long-term Treasury bonds gained 33.9%. In 2018 and 2022 both fell. Of the three years in which shares fell between 2006 and 2025, bonds rose in one. Correlation, which measures whether two prices move together, was -0.30 before 2022 and 0.56 from 2022.
What happened to bonds in 2022?
Long-term Treasury bonds lost 31.2% in 2022, while shares lost 18.2%. The US central bank raised interest rates through the year. Rising rates push down the prices of existing bonds. The bond fund's fall from its August 2020 high reached 48.3%. On 26 October 2026 it was still 42.1% below that high.
Are bonds safer than stocks?
Not by every measure. Long-term Treasury bonds swung less from day to day, at 14.73% a year against 19.17% for shares. Their biggest fall from a previous high was still 48.3%, against 55.2% for shares. That fall lasted longer. Shorter-dated bonds were not tested.
DISCLAIMER: This article is strictly educational. Any information or analysis in this note is not an offer to sell or the solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice and nor should it be relied upon to make investment decisions. Any opinions, analyses, or probabilities expressed in this note are those of the author as of the note's date of publication and are subject to change without notice.
