Stocks vs Bonds
How far have stocks pulled ahead of Treasuries—and what has a lead like this meant for the next decade?
A rising line means stocks are growing wealth faster; a falling line means Treasuries are. Dividends and coupons count. Both return series are estimates, with the Treasury side modeled from yields.
S&P 500: 15.4% a year.
Modeled Treasury: 0.1% a year.
Same endpoint: Oct 22, 2026. Points are the difference between the two annualized returns.
- Leader under the 25% rule
- Stocks
- Retrospective turn
- Mar 23, 2020
- Confirmed later
- Apr 14, 2020
- Ratio vs record
- −4.5%
The separate monthly record ends Jun 2026. Its ten-year stock lead was +15.2 pts a year, wider than 97% of prior readings.
The long contest, since 1970
Stocks / Treasuries, with income reinvested
Jan 2, 1970 = 100 · log scaleStocks / Treasuries, with income reinvested
Jan 2, 1970 = 100 · log scaleOct 1974 → Oct 1987
+195%
Ratio change over 13 years
Jan 2000 → Mar 2009
−73%
Ratio change over 9.2 years
Mar 2020 → Oct 2026
+363%
Ratio change over 6.5 years
A 100% rise means stock wealth doubled relative to Treasury wealth. It does not mean stocks earned 100 percentage points more. The comparison windows span intervening reversals.
See the S&P 500 price alongside this history
A smaller lead can still be a stock win
After unusually strong decades, stocks often kept outperforming. Their advantage usually shrank. Those are different questions, and the historical record answers both.
of wide-lead months were followed by another decade of stock outperformance
were followed by a narrower annual return gap than in the preceding decade
The wide-lead group's median trailing gap was +12.4 pts a year; its median following gap was +6.5 pts a year. Stock wins were less common than the 85% rate across all eligible months. A wide lead alone was an unreliable reason to expect bonds to take over.
| Prior stock lead | Past 10y gap | Next 10y gap | Stocks win | Months |
|---|---|---|---|---|
| Widest fifth | +12.4 pts | +6.5 pts | 75% | 292 |
| All eligible months | +3.7 pts | +4.5 pts | 85% | 1266 |
| Narrowest fifth | −1.8 pts | +6.5 pts | 100% | 229 |
Gap columns are cohort medians of stock CAGR minus Treasury CAGR, in percentage points per year. The narrowing rate compares past and future gaps within each observation; it is not calculated by subtracting the two medians.
How much separate history is there?
Those 292 wide-lead months contain only 5 non-overlapping forward decades under the selection below. Stocks won 4 of them; the median next-decade gap was +5.4 pts a year. That is a small sample, sensitive to which starting months are selected.
| Forward decade | Past gap | Next gap | Winner |
|---|---|---|---|
| Dec 1924 → Dec 1934 | +5.3 pts | +0.2 pts | Stocks |
| May 1942 → May 1952 | +6.0 pts | +15.8 pts | Stocks |
| May 1952 → May 1962 | +15.8 pts | +12.4 pts | Stocks |
| May 1962 → May 1972 | +12.4 pts | +5.4 pts | Stocks |
| Apr 1998 → Apr 2008 | +9.0 pts | −2.7 pts | Bonds |
Selection: take the earliest qualifying month, then the next at least ten years later. Forward windows do not overlap; their economic conditions and trailing windows can. Non-overlap does not establish statistical independence.
Study rules, the five-year check and the longer record
Each month's trailing ten-year gap is ranked against earlier gaps, with at least 240 prior months required. The top fifth is a wide stock lead; the bottom fifth is a narrow lead, which can still be positive. Outcomes need a complete forward window. The ten-year sample therefore ends June 2016.
At five years, the wide-lead group's median forward gap was +8.5 pts a year versus +6.2 pts across all eligible months; stocks won 79% versus 77%. Those 323 overlapping months reduce to 9 non-overlapping windows, whose median gap was +11.0 pts.
Ranks use only earlier observations, but the underlying inputs are revised historical series without release lags. The study describes history; it does not reconstruct the information investors had at each release. The monthly record uses monthly average prices and yields; it is kept separate from the daily close series.
A turn is visible before it is confirmed
The current stocks stretch traces back to Mar 23, 2020. The 25% reversal rule confirmed it on Apr 14, 2020. The earlier date identifies a low or high in hindsight. Investors could only observe the confirmation later.
The ratio is −4.5% from this stretch's high on Oct 6, 2026. A 25% fall from the high would confirm a reversal—−21.5% from the latest level. This describes relative leadership; it is not an allocation rule.
Explore all 24 leadership stretches since 1962
All returns below are cumulative across the displayed dates, including short stretches. A closed stretch ends at a retrospective extreme; its end was confirmed later. The open stretch ends at the latest observation. Gains measure wealth in each asset; ratio change measures one relative to the other.
| Pivot → end | Leader | Start confirmed | End confirmed | Length | Stocks | Treasury | Ratio |
|---|---|---|---|---|---|---|---|
| Mar 23, 2020 → latest | Stocks · open | Apr 14, 2020 | — | 6.6 years | +276.6% | −14.9% | +342.6% |
| Jan 17, 2020 → Mar 23, 2020 | Bonds | Mar 9, 2020 | Apr 14, 2020 | 2 months | −32.6% | +10.8% | −39.1% |
| Oct 3, 2011 → Jan 17, 2020 | Stocks | Feb 8, 2012 | Mar 9, 2020 | 8.3 years | +257.6% | +19.2% | +199.9% |
| Feb 18, 2011 → Oct 3, 2011 | Bonds | Aug 8, 2011 | Feb 8, 2012 | 7 months | −17.2% | +18.9% | −30.3% |
| Mar 9, 2009 → Feb 18, 2011 | Stocks | Apr 9, 2009 | Aug 8, 2011 | 1.9 years | +107.2% | +0.1% | +107.0% |
| Jul 13, 2007 → Mar 9, 2009 | Bonds | Jan 22, 2008 | Apr 9, 2009 | 1.7 years | −54.7% | +27.1% | −64.4% |
| Oct 9, 2002 → Jul 13, 2007 | Stocks | Nov 21, 2002 | Jan 22, 2008 | 4.8 years | +117.1% | +8.7% | +99.7% |
| Mar 19, 2002 → Oct 9, 2002 | Bonds | Jul 10, 2002 | Nov 21, 2002 | 7 months | −33.0% | +17.7% | −43.1% |
| Sep 21, 2001 → Mar 19, 2002 | Stocks | Mar 11, 2002 | Jul 10, 2002 | 6 months | +22.0% | −2.6% | +25.2% |
| Jan 14, 2000 → Sep 21, 2001 | Bonds | Dec 20, 2000 | Mar 11, 2002 | 1.7 years | −32.7% | +27.6% | −47.3% |
| Oct 7, 1998 → Jan 14, 2000 | Stocks | Nov 23, 1998 | Dec 20, 2000 | 15 months | +53.4% | −10.5% | +71.3% |
| Jul 17, 1998 → Oct 7, 1998 | Bonds | Oct 5, 1998 | Nov 23, 1998 | 3 months | −17.9% | +10.8% | −25.9% |
| Dec 4, 1987 → Jul 17, 1998 | Stocks | Jan 31, 1989 | Oct 5, 1998 | 10.6 years | +611.7% | +172.3% | +161.3% |
| Oct 5, 1987 → Dec 4, 1987 | Bonds | Oct 19, 1987 | Jan 31, 1989 | 2 months | −31.4% | +6.5% | −35.5% |
| Aug 16, 1982 → Oct 5, 1987 | Stocks | Apr 15, 1983 | Oct 19, 1987 | 5.2 years | +287.6% | +104.9% | +89.2% |
| Aug 5, 1981 → Aug 16, 1982 | Bonds | Mar 3, 1982 | Apr 15, 1983 | 12 months | −16.8% | +27.2% | −34.6% |
| May 6, 1980 → Aug 5, 1981 | Stocks | Aug 11, 1980 | Mar 3, 1982 | 15 months | +32.6% | −13.0% | +52.4% |
| Feb 26, 1980 → May 6, 1980 | Bonds | Apr 21, 1980 | Aug 11, 1980 | 2 months | −5.8% | +27.8% | −26.3% |
| Oct 4, 1974 → Feb 26, 1980 | Stocks | Mar 3, 1975 | Apr 21, 1980 | 5.4 years | +134.5% | +10.8% | +111.7% |
| Jan 11, 1973 → Oct 4, 1974 | Bonds | May 22, 1974 | Mar 3, 1975 | 1.8 years | −44.9% | +1.1% | −45.5% |
| Jul 7, 1970 → Jan 11, 1973 | Stocks | Apr 14, 1971 | May 22, 1974 | 2.5 years | +82.7% | +26.2% | +44.8% |
| Dec 3, 1968 → Jul 7, 1970 | Bonds | May 20, 1970 | Apr 14, 1971 | 1.6 years | −30.4% | +0.1% | −30.5% |
| Jun 26, 1962 → Dec 3, 1968 | Stocks | Jan 28, 1963 | May 20, 1970 | 6.4 years | +152.8% | +15.1% | +119.6% |
| Jan 2, 1962 → Jun 26, 1962 | Bonds | Jun 21, 1962 | Jan 28, 1963 | 6 months | −25.2% | +2.6% | −27.1% |
Sources, assumptions and the reference chart
The Treasury line estimates returns from yields. It does not track a fund's observed total return. That distinction matters most when interpreting exact levels across many decades.
How the estimates are calculated and dated
Stock returns. Daily S&P 500 closes plus Shiller's annual dividend rate accrued over calendar days and reinvested. Dividend observations end June 2026; the last known dividend yield estimates the following 4 months. This is not an official S&P total-return index.
Treasury returns. At each observation, a par ten-year note paying the previous yield is repriced using twenty semiannual coupon periods at the current yield, then coupon accrual is added. Proceeds are treated as reinvested. This approximates a constant-maturity roll: it does not age the note between observations, use the full yield curve, or deduct trading costs and fees. DGS10 itself measures a yield. Bond price changes must be estimated.
Dates. Stock closes are available through Oct 23, 2026; Treasury yields through Oct 22, 2026. The headline and daily chart stop at their latest common date, Oct 22, 2026. Inside the record, stock sessions without a new Treasury yield use the latest prior bond observation. The separate monthly study ends June 2026.
Comparison with IEF. From Jul 30, 2002 to Oct 22, 2026, monthly log returns correlated 0.990. The model earned 2.96% a year against IEF's 3.46%; annualized monthly volatility was 7.58% against 6.64%. IEF holds a 7–10-year basket, so it is a useful cross-check with a different maturity profile. A high correlation does not verify the model's accumulated wealth level.
Sources: S&P 500 closes, FRED DGS10, Robert Shiller's historical series, IEF, and downloadable calculations and chart data.
Can this reproduce the “adjusted for inflation” reference chart?
The reference image names the S&P 500, the ten-year Treasury and inflation, but supplies no formula or numeric vertical axis. We tested 28 plausible constructions. None verified both the shape and its annotated gains. The page adopts its long view and dated comparisons without claiming to reproduce its underlying measure.
The closest visual candidate was S&P 500 price × ten-year yield ÷ CPI. Multiplying by a yield does not measure bond investment performance. It also failed to reproduce the image's gain labels, so it is not the headline series here.
If both wealth indexes are divided by the same CPI, inflation cancels: (stock wealth ÷ CPI) ÷ (bond wealth ÷ CPI) = stock wealth ÷ bond wealth. Dividing the ratio itself by CPI produces a different quantity. This says nothing definitive about the unidentified formula in the image.
CPI: BLS via FRED, through September 2026.
Does a wide stock lead predict a bond rally?
After the widest fifth of stock leads, stocks outperformed over the next 10 years in 75% of months, versus 85% across all eligible months. Yet the annual return gap narrowed in 74% of those wide-lead months. These are overlapping historical observations, not independent trials or a forecast. Relative outperformance does not imply a positive absolute return for either asset.
What does the stocks-to-bonds ratio measure?
Estimated wealth from stocks with dividends reinvested divided by modeled Treasury wealth with coupons reinvested. A rising ratio means stocks outperform; it does not guarantee that either investment gains in absolute terms. The daily chart begins at 100 in January 1970.
Does inflation change relative wealth?
Dividing both wealth indexes by the same consumer price index leaves their ratio unchanged. Dividing the ratio itself by CPI produces a different measure. A stock-price-to-yield formula is also different: a Treasury yield is not a bond total-return index.
Are historical turns available as trading signals?
A high or low is labeled retrospectively after a 25% fall from the high or a 25% rise from the low. The page shows the later confirmation date separately. Historical pivots and their returns should not be read as results achievable by trading on the pivot dates.
S&P 500: 15.4% a year.
Modeled Treasury: 0.1% a year.
Same endpoint: Oct 22, 2026. Points are the difference between the two annualized returns.
- Leader under the 25% rule
- Stocks
- Retrospective turn
- Mar 23, 2020
- Confirmed later
- Apr 14, 2020
- Ratio vs record
- −4.5%
The separate monthly record ends Jun 2026. Its ten-year stock lead was +15.2 pts a year, wider than 97% of prior readings.