Guide

Stocks vs bonds: what each did in every crash, tested

Over the long run stocks grew far faster than bonds, but bonds did not cushion every crash: from 10 April 2007 to 20 November 2026, the stock fund SPY grew 10.3% a year against 2.9% for the bond fund BND. BND returned +6.9% in 2008, when SPY returned −36.8%, but −13.1% in 2022, when stocks and bonds fell together.

What stocks and bonds are

A stock is a small share of a company. Its value rises and falls with the company's profits and with what other investors will pay for them, so it can grow a lot or fall a long way. A bond is a loan: you lend money to a government or a company, and it pays you interest and returns the loan on a set date. Bonds usually swing less, because the payments are fixed in advance.

Most people own both through funds that hold hundreds or thousands of them. The trap is that "bonds" is not one thing. A bond fund's price moves opposite to interest rates: when rates rise, older bonds paying less interest are worth less. The longer a fund's bonds have until they are repaid, the bigger that effect. This sensitivity to rates is called duration, and it is why two bond funds can behave very differently in the same year.

Stocks and bonds side by side

Stocks

  • A share of a company
  • Paid from profits, nothing fixed
  • Can grow a lot
  • Can fall a long way

Bonds

  • A loan to a government or company
  • Fixed interest and repayment
  • Price moves opposite to rates
  • Longer bonds move more

The follow-along example

We test three funds on their own and one mix, each starting with $10,000, dividends and interest reinvested, from 10 April 2007 (the first day all three funds have prices) to 20 November 2026.

Follow along: the four portfolios

Portfolio What it holds Rule
SPY US large-company stocks (the S&P 500 fund) Held; our benchmark
BND The whole US investment-grade bond market, medium duration Held
TLT US government bonds repaid in 20 years or more, long duration Held
SPY 60 / BND 40 60% stocks, 40% bonds Reset to 60/40 once a year

60/40 means six-tenths of the money in stocks and four-tenths in bonds. Rebalancing means selling some of whatever has grown and buying what has lagged, to get back to those weights. The benchmark is the yardstick: everything is measured against SPY.

How we tested this

DataDaily prices adjusted for splits, dividends and interest (total return) for SPY, BND and TLT, from EODHD.
Period10 April 2007 to 20 November 2026, the longest period every holding has prices for
BasketsSPY 60 / BND 40: BND 40%, SPY 60%, rebalanced yearly; BND (bonds): BND 100%, held throughout; TLT (long Treasuries): TLT 100%, held throughout; SPY (stocks): SPY 100%, held throughout (benchmark)
How we testedEach portfolio starts with $10,000 on the first day all three funds have prices. SPY, BND and TLT are held with no rebalancing; SPY 60 / BND 40 is reset to its weights once a year. SPY is the benchmark. Rolling windows start once a month; the two halves split the history at its midpoint. Sharpe and Sortino use a 0% risk-free rate.
Not includedTrading costs and taxes; fund fees are already in each fund's price. Cash, short-term and international bonds, and markets before 2007, including the 2000 to 2002 bear market and the 1970s.

Step 1: Look at the worst fall first

Start with the number that decides whether you can stick with a portfolio: its max drawdown, the largest fall from a high to a later low (max drawdown explained). Long government bonds fell almost as far as stocks, and the total bond fund's worst fall came in the bond crash after 2020, not in 2008.

SPY 60 / BND 40BND (bonds)TLT (long Treasuries)SPY (stocks)
Drawdown from the previous high, full period. Worst falls: SPY 60 / BND 40 33.7%; BND (bonds) 18.8%; TLT (long Treasuries) 48.4%; SPY (stocks) 55.2%-40%-20%0%2008201020122014201620182020202220242026
  • SPY 60 / BND 40 worst fall 33.7% (9 October 2007 to 9 March 2009), longest time below a previous high 3.1 years
  • BND (bonds) worst fall 18.8% (6 August 2020 to 20 October 2022), longest time below a previous high 6.3 years
  • TLT (long Treasuries) worst fall 48.4% (4 August 2020 to 19 October 2023), longest time below a previous high 6.3 years
  • SPY (stocks) worst fall 55.2% (9 October 2007 to 9 March 2009), longest time below a previous high 4.9 years

Full period: Indices Lab backtest, total return, mixes rebalanced yearly, single funds held throughout, 10 April 2007 to 20 November 2026. Hypothetical results: past performance does not predict future returns.

Follow along: the deepest fall and the longest wait

Portfolio Max drawdown Longest time below a previous high
SPY 55.2% 4.9 years
BND 18.8% 6.3 years
TLT 48.4% 6.3 years
SPY 60 / BND 40 33.7% 3.1 years

The surprise is the right-hand column. SPY's worst fall was the deepest, but both bond funds spent longer below their old high than SPY ever did, because their long stretch began in 2020 and the interest they pay rebuilds value slowly. A smaller fall is not the same as a shorter wait.

Step 2: Check each crash year on its own

"Bonds go up when stocks go down" is the most repeated line about the two. Test it year by year, because it depends on why stocks are falling.

SPY 60 / BND 40BND (bonds)TLT (long Treasuries)SPY (stocks)
Calendar-year returns, SPY 60 / BND 40 best +22.3% (2019), worst -19.3% (2008); BND (bonds) best +8.8% (2019), worst -13.1% (2022); TLT (long Treasuries) best +34.0% (2011), worst -31.2% (2022); SPY (stocks) best +32.3% (2013), worst -36.8% (2008)-40%-20%0%+20%200820102012201420162018202020222024
200820092010201120122013201420152016201720182019202020212022202320242025
SPY 60 / BND 40−19.3%+17.3%+11.5%+4.3%+11.1%+18.5%+10.4%+1.0%+8.2%+14.5%−2.8%+22.3%+14.0%+16.4%−16.2%+18.0%+15.5%+13.5%
BND (bonds)+6.9%+3.6%+6.2%+7.9%+3.9%−2.1%+5.8%+0.6%+2.5%+3.6%−0.1%+8.8%+7.5%−2.0%−13.1%+5.7%+1.4%+7.1%
TLT (long Treasuries)+33.9%−21.8%+9.0%+34.0%+2.6%−13.4%+27.3%−1.8%+1.2%+9.2%−1.6%+14.1%+18.2%−4.6%−31.2%+2.8%−8.1%+4.2%
SPY (stocks)−36.8%+26.4%+15.1%+1.9%+16.0%+32.3%+13.5%+1.2%+12.0%+21.7%−4.6%+31.2%+18.3%+28.7%−18.2%+26.2%+24.9%+17.7%

Full calendar years only. Full period: Indices Lab backtest, total return, mixes rebalanced yearly, single funds held throughout, 10 April 2007 to 20 November 2026. Hypothetical results: past performance does not predict future returns.

Follow along: the years stocks fell

Year SPY BND TLT SPY 60 / BND 40
2008 −36.8% +6.9% +33.9% −19.3%
2018 −4.6% −0.1% −1.6% −2.8%
2022 −18.2% −13.1% −31.2% −16.2%

In 2008 the economy was in trouble, interest rates were cut, and bonds rose while stocks fell: TLT, the long fund, did best. In 2022 the problem was inflation, rates went up fast, and everything fell in the same year. TLT, the best cushion in 2008, lost the most of the three funds in 2022.

So the cushion works when stocks fall because growth is weak and rates come down, and fails when stocks fall because rates go up. Long-duration bonds amplify both cases.

Step 3: Compare growth over the whole period

Now the long run. Read growth as CAGR, the compound annual growth rate: the steady yearly rate that turns the start value into the end value (how CAGR works).

SPY 60 / BND 40BND (bonds)TLT (long Treasuries)SPY (stocks)
Growth of $10,000, full period: SPY 60 / BND 40 $42,993, BND (bonds) $17,601, TLT (long Treasuries) $17,247, SPY (stocks) $68,107$20k$40k$60k$80k2008201020122014201620182020202220242026
  • SPY 60 / BND 40 $42,993
  • BND (bonds) $17,601
  • TLT (long Treasuries) $17,247
  • SPY (stocks) $68,107

Full period: Indices Lab backtest, total return, mixes rebalanced yearly, single funds held throughout, 10 April 2007 to 20 November 2026. Hypothetical results: past performance does not predict future returns.

Follow along: the whole period

Portfolio CAGR Gap to SPY, per year End value
SPY 10.3% $68,107
BND 2.9% −7.4 points $17,601
TLT 2.8% −7.5 points $17,247
SPY 60 / BND 40 7.7% −2.6 points $42,993

Stocks ended far ahead. TLT fell almost as far as stocks in Step 1, yet grew no faster than the much calmer BND: it took stock-sized risk for a bond-sized return.

Clone this 60/40 example Opens the builder with BND 40%, SPY 60% against SPY (stocks). Free account, no card.

Step 4: Check every start date

One start date can flatter or punish anything. A rolling window test takes every possible 5-year or 10-year stretch, one per month, and counts how often each portfolio finished ahead of SPY. The chart shows the 60/40 mix against SPY in every 10-year window.

SPY 60 / BND 40 aheadSPY (stocks) ahead
SPY 60 / BND 40 minus SPY (stocks), annualised return over every rolling 10-year window: SPY 60 / BND 40 ahead in 0.0% of 116 windows-4 pts-2 pts0 pts201820192020202120222023202420252026

SPY 60 / BND 40 finished ahead of SPY (stocks) in 0.0% of 116 rolling 10-year windows (window ends by month).

Full period: Indices Lab backtest, total return, mixes rebalanced yearly, single funds held throughout, 10 April 2007 to 20 November 2026. Hypothetical results: past performance does not predict future returns.

Follow along: how often bonds beat stocks

Portfolio Ahead of SPY in 5-year windows Ahead in 10-year windows Last 5-year window it won ended
BND 7.4% of 176 0.0% of 116 May 2013
TLT 9.7% of 176 1.7% of 116 March 2020
SPY 60 / BND 40 8.5% of 176 0.0% of 116 June 2013

Almost every 5-year window that bonds won started between 2007 and mid-2008, near the old stock high, and ran through the 2008 crash. TLT also won two that ended in a stock sell-off, the last in the 2020 crash. In calendar years they did better than SPY more often: BND in 22.2% of full years.

Two traps here. The 10-year windows overlap heavily, so treat them as one long result, not a hundred tests. And stocks rose strongly for most of this history; a long stock slump such as 2000 to 2002 is not in the test.

Step 5: Split the history in two

Bonds' returns depend on where interest rates start and where they go. Split the test at its midpoint and the two halves tell opposite stories.

Follow along: growth per year in each half

Portfolio First half Second half
SPY 7.0% 13.6%
BND 4.2% 1.7%
TLT 6.8% -1.0%
SPY 60 / BND 40 6.5% 9.0%

In the first half, with rates falling, long government bonds kept pace with stocks and finished slightly ahead. In the second half, with rates ending well above where they started, TLT lost money over almost ten years.

In the first half, TLT and SPY finished close together:

SPY 60 / BND 40BND (bonds)TLT (long Treasuries)SPY (stocks)
Growth of $10,000, first half: SPY 60 / BND 40 $18,458, BND (bonds) $14,926, TLT (long Treasuries) $18,985, SPY (stocks) $19,451$5k$10k$15k$20k2008200920102011201220132014201520162017
  • SPY 60 / BND 40 $18,458
  • BND (bonds) $14,926
  • TLT (long Treasuries) $18,985
  • SPY (stocks) $19,451

First half: Indices Lab backtest, total return, mixes rebalanced yearly, single funds held throughout, 10 April 2007 to 26 January 2017. Hypothetical results: past performance does not predict future returns.

The lesson: a bond fund's past return mostly tells you what rates did. When you judge bonds on a backtest, ask what rates were at the start and what they are now.

Step 6: Mix them and measure the ride

The case for holding both is not that bonds earn more. It is that the mix can give more return for each unit of swing. Volatility measures how much a portfolio swings up and down, as a yearly percentage. The Sharpe ratio divides return by volatility, so higher means more return per unit of swing (the risk ratios compared).

BasketYearly returnVolatilitySharpeWorst drop$10,000 became
SPY 60 / BND 407.7%11.5%0.7133.7%$42,993
BND (bonds)2.9%5.2%0.5818.8%$17,601
TLT (long Treasuries)2.8%15.1%0.2648.4%$17,247
SPY (stocks)10.3%19.6%0.6055.2%$68,107

Full period: Indices Lab backtest, total return, mixes rebalanced yearly, single funds held throughout, 10 April 2007 to 20 November 2026. Hypothetical results: past performance does not predict future returns. Sharpe uses a 0% risk-free rate.

Follow along: return for the risk taken

Portfolio Volatility Sharpe Sortino
SPY 19.6% 0.60 0.84
BND 5.2% 0.58 0.83
TLT 15.1% 0.26 0.37
SPY 60 / BND 40 11.5% 0.71 1.00

The 60/40 mix had a higher Sharpe ratio than SPY or BND on its own: holding both gave more return per unit of swing than either part. Its growth gap to SPY was −2.6 points a year: the price of a worst fall of 33.7% instead of 55.2%.

The Sortino ratio is the same idea counting only downward swings. For more mixes, including three-fund and All Weather portfolios, see lazy portfolios backtested.

What this example can't show

It starts in 2007, when BND began trading, so it misses the 2000 to 2002 bear market and the high-inflation 1970s. It covers mostly falling rates, then one sharp rise. It tests US funds only, and not cash or short-term bonds. And it uses no costs or taxes: bond interest is often taxed more heavily than stock gains.

Before you choose a stock and bond mix

  • Look at the worst fall and the longest wait
  • Check each crash year on its own
  • Know your bond fund's duration
  • Test every start date, not one
  • Split the history in two
  • Ask what rates were at the start
  • Compare return per unit of swing

Test your own stock and bond mix over the same dates, then change the weights and the bond fund and see what moves.

Backtest your own stock and bond mix Opens the builder with BND 40%, SPY 60% against SPY (stocks). Free account, no card.

Questions

Are bonds safer than stocks?

They swing less, but not always by as much as people expect. From 10 April 2007, the total bond fund BND had a worst fall of 18.8% against 55.2% for SPY, while the long government bond fund TLT fell 48.4%. How far a bond fund can fall depends mostly on its duration.

Do bonds go up when stocks go down?

Sometimes. In 2008, when rates were cut, BND returned +6.9% and TLT +33.9% while SPY returned −36.8%. In 2022, when rates rose to fight inflation, all three fell in the same year.

Why did bonds fall in 2022?

Interest rates rose quickly, and a bond fund's price moves opposite to rates: older bonds paying less interest became worth less. Funds holding long bonds fell most; TLT returned −31.2% that year.

Do stocks always beat bonds over 10 years?

In this test, BND finished ahead of SPY in 0.0% of 116 ten-year windows. The test starts in 2007, so it misses long stock slumps such as 2000 to 2002.

What is a 60/40 portfolio?

A mix with 60% in stocks and 40% in bonds, reset to those weights on a schedule. Here SPY 60 / BND 40, reset once a year, grew 7.7% a year with a worst fall of 33.7%.

Build your index Opens the builder with BND 40%, SPY 60% against SPY (stocks). Free account, no card.

About this page. Figures on this page come from Indices Lab’s own data, powered by EODHD, and every number is recalculated after each market close; the text is reviewed by a person before it is published. Indices Lab is a research and educational tool: everything here is hypothetical and informational, not investment advice and not a recommendation to buy or sell any security. Past performance does not predict future results. See our methodology.