In 2008, with dividends counted, SPY (the S&P 500 ETF) lost 36.8% while TLT (the long-term US Treasury bond ETF) gained 33.9%. In 2022, SPY lost 18.2% and TLT lost 31.2%. The two funds had gone from moving in opposite directions to moving together.
Correlation is a number from -1 to +1 that shows whether two investments move together or apart. When it is below zero, they tend to move in opposite directions. This guide measures it for SPY and TLT from 26 July 2002 to 26 October 2026, with IEF (the 7-to-10-year Treasury bond ETF) as a check. It shows when the correlation went from below zero to above zero.
What is correlation in investing? Correlation is a number from -1 to +1 that measures how closely the returns of two investments move together. At +1 they move in step. At -1 they move in opposite directions. Near 0 there is no steady link.
From 6 July 2007 to 9 November 2022, the correlation of SPY and TLT over the previous 252 trading days (about one year) was below zero on every day. It has been above zero on every day since 9 April 2025.
Our calculation of a mix of half SPY and half TLT, reset to half and half each day, lost 4.9% in 2008, when SPY lost 36.8%. The same mix lost 24.1% in 2022.
Correlation describes one period. The result changes if you use weekly returns instead of daily returns, or a different number of past days.
What does correlation mean in investing?
A return is the percentage change in a price over a period, here one trading day. Correlation compares two lists of returns, day by day. If both rise on the same days and fall on the same days, it is close to +1. If one rises when the other falls, it is close to -1.
The worst day for SPY in the sample was 16 March 2020, when it lost 10.94%. TLT gained 6.48% that day. Correlation summarises thousands of such days.
The sample has 6,102 trading days of prices, from 26 July 2002, which give 6,101 daily returns, the first on 29 July 2002.
SPY holds the S&P 500, the index of large US companies. TLT holds US Treasury bonds with over 20 years left. IEF holds Treasuries with 7 to 10 years left. Our explainer on bond prices and yields covers them.
Over the whole sample, the daily correlation of SPY with TLT is -0.30. With IEF it is -0.29. TLT and IEF move almost together, at 0.91.
Size matters as well as sign. A reading of 0.91 is close to moving in step. One near 0.30 is a loose link.
Were SPY and TLT always negatively correlated?
No. Chart 1 plots the correlation of daily SPY and TLT returns over the previous 252 trading days, about one year. We call this the one-year line.
Of the 5,850 values from 28 July 2003 to 26 October 2026, the lowest was -0.77 on 8 August 2012. The highest was +0.32 on 13 October 2026.

Source: YX Insights price data
The line sat above zero on 322 of the first 992 days, from 28 July 2003 to 5 July 2007. It fell below zero on 6 July 2007 and stayed there on all 3,866 trading days to 9 November 2022.
How did the correlation vary by calendar year?
Chart 2 shows calendar years. Seventeen of the 19 full years from 2003 to 2021 were negative. The exceptions were 2005 at +0.02 and 2006 at +0.05. The most negative year was 2011, at -0.71.

Source: YX Insights price data
The bars from 2022 onward are all above zero, while those from 2007 to 2021 are all below it.
What did the correlation do after 2021?
The one-year line in Chart 1 crossed above zero on 10 November 2022. It stayed above zero until 3 April 2025. It dipped below zero from 4 to 8 April 2025.
On 9 April 2025 it returned above zero and stayed there, standing at +0.29 on 26 October 2026.
The calendar-year bars agree: the correlation was +0.08 in 2022, +0.13 in 2023, +0.06 in 2024, +0.10 in 2025 and +0.30 in 2026 to 26 October. These readings are small. The sign is what moved.
Interest rates rose sharply in 2022. The Federal Reserve (the Fed) sets a target range for the Federal Funds Rate, the rate banks charge each other for overnight loans. The upper limit of that range (FRED series DFEDTARU) went from 0.25% on 31 December 2021 to 4.50%, the rate in force on 31 December 2022, a Saturday.
The 10-year Treasury yield, the annual return the bond pays at its current price, went from 1.52% on 31 December 2021 to 3.88% on 30 December 2022. Bond prices fall when yields rise. Our explainer on the 2022 rate hikes gives the path.
This test does not measure whether rates caused the change in correlation.
Why did the sign matter for a stock and bond mix?
Combining two investments that do not move together lowers the swings of the mix. Volatility measures them: the standard deviation of daily returns, scaled to a year. Our explainer on volatility defines it.
We built a 50/50 mix of SPY and TLT, reset to half and half every day. Chart 3 compares our calculation of the mix with each fund in 2008 and 2022. TLT gained 33.9% in 2008 and lost 31.2% in 2022.

Source: YX Insights price data
In 2008 our mix lost 4.9%, against 36.8% for SPY. In 2022 our mix lost 24.1%, 5.9 percentage points worse than SPY, because TLT lost more than SPY.
Volatility fell by less after 2021. From 29 July 2002 to 31 December 2021, our mix had 9.27% a year, against 19.13% for SPY. From 3 January 2022 it had 12.26%, against 17.24% for SPY.
Our guides to stocks against bonds and the 60/40 portfolio cover the full window.
What does a correlation number leave out?
First, it says nothing about size. TLT and IEF have a correlation of 0.91, yet in 2022 TLT lost 31.2% and IEF lost 15.2%, so TLT lost 2.06 times as much. Our comparison of IEF and TLT explains the gap.
Second, it depends on the return interval, the length of each return, such as one day or one week. We repeated the test on weekly returns, from each Friday's last close. The two periods split at the end of 2021 still flip, as the table shows.
Correlation Of SPY And TLT Returns Before And After 2021, Daily And Weekly
Period | Daily returns | SPY and TLT, daily | SPY and IEF, daily | SPY and TLT, weekly |
|---|---|---|---|---|
29 July 2002 to 31 December 2021 | 4,893 | -0.40 | -0.40 | -0.35 |
3 January 2022 to 26 October 2026 | 1,208 | +0.11 | +0.11 | +0.14 |
Whole window | 6,101 | -0.30 | -0.29 | -0.25 |
Correlation of returns, with dividends counted. Weekly returns run Friday close to Friday close, 1,014 weeks to 31 December 2021 and 251 weeks from 2022. Source: YX Insights price data
Daily and weekly returns both give a negative number before 2022 and a positive one after. Over a single year the interval can flip the sign. Weekly returns for 2025 alone give -0.25, while daily returns give +0.10.
Third, it depends on the window, the number of past days or weeks the number uses.
Over 504 trading days, every value from 2008 to 2021 is negative and 99.2% of values since 2023 are positive. The turn shows in every version. It is sharpest with 252 days of daily returns.
What does this test assume?
The test covers three US-listed funds over one window. The numbers use adjusted closing prices, which count dividends. They ignore costs and tax. Chart 2 starts at 2003 because 2002 is a part year. Its 2026 bar runs to 26 October.
The mix, the window and the split date are our choices.
The one-year line for SPY and TLT was below zero on every day from 6 July 2007 to 9 November 2022. It has been above zero since 9 April 2025.
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 correlation in investing
What does a correlation of 0.5 mean in investing?
A correlation of 0.5 means two investments tended to move in the same direction, though loosely. It sits halfway between 0 (no steady link) and +1 (perfect step). For example, no 252-day reading for SPY and TLT reached 0.5. The highest was +0.32 on 13 October 2026.
Is a negative correlation good for a portfolio?
It lowered the swings of a mix. From 2002 to 2021, our 50/50 mix of SPY and TLT had volatility of 9.27% a year, against 19.13% for SPY. Volatility is the yearly size of daily swings. The reduction was smaller from 2022: 12.26% against 17.24% for SPY.
What happened to stocks and bonds in 2022?
Bonds and stocks both fell in 2022. TLT lost 31.2% while the 10-year Treasury yield rose from 1.52% to 3.88%. SPY lost 18.2%. Our data show the losses and the yield change, not the cause of the stock fall. Bond prices fall when yields rise.
How is correlation calculated?
Correlation takes each fund's average daily return. It measures how far each day sits from that average. It then compares those distances across the two funds. A result of +1 means the two move in exact proportion. A result of -1 means they move in exact opposite proportion.
Do stocks and bonds still move in opposite directions?
Not on the one-year measure. SPY and TLT have had a positive 252-day correlation on every day since 9 April 2025, standing at +0.29 on 26 October 2026. The calendar-year figure for 2026 to date is +0.30. Over the whole window since 2002 the daily figure is still negative, at -0.30.
DISCLAIMER: This article is intended for educational purposes only. Any information or analysis in this note does not constitute an offer to sell or a solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice, 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.