How bond ETFs move with interest rates

What happens to TLT, IEF and SHY when rates rise or fall, when the Fed buys or sells bonds, if rates go negative, and when stocks look expensive, with a look-ahead where you place Fed rate cuts and hikes on the chart. Built on iShares’ live figures and a century of history.

A bond ETF’s price moves opposite to interest rates. How far it moves is its duration: the rough percent the price changes for each 1 point change in rates. Convexity bends that line in the holder’s favour, so big falls in rates help a little more than big rises hurt. The yield to maturity is what the bonds pay if rates never change, before the 0.15% fund fee.

SHY, IEF and TLT: iShares portfolio characteristics and what a 1 point rate rise does
Fund	Average maturity	Effective duration	Convexity	Yield to maturity	30-day SEC yield	If rates rise 1 point today	Time to earn it back	Actual volatility (3 years)
SHY (opens in a new tab)	1.89 yrs	1.78 yrs	0.04	4.84%	4.50%	−1.8%	4 months	1.52%
IEF (opens in a new tab)	8.42 yrs	6.84 yrs	0.56	5.22%	4.91%	−6.6%	1.2 years	6.57%
TLT (opens in a new tab)	26.03 yrs	14.74 yrs	3.09	5.65%	5.49%	−13.2%	2.3 years	13.80%

Loading today’s figures from iShares; showing those from Sep 29, 2026 until then. “Time to earn it back” is how long the fund’s new, higher yield takes to repay the price drop.

Fund

Amount invested

Rates move byDrawn both ways: rates up by this much, and down by this much.

1.00 points

OverThe move is spread evenly over this time. 0 is an overnight jump.

2 years

You hold the fund for

10 years

Rates up 1 pointRates unchangedRates down 1 point

Value at each year
Year	Rates up	Unchanged	Rates down
0	$10.0K	$10.0K	$10.0K
1	$9.8K	$10.6K	$11.3K
2	$9.7K	$11.2K	$12.8K
3	$10.4K	$11.8K	$13.4K
4	$11.1K	$12.5K	$14.0K
5	$11.8K	$13.2K	$14.7K

If rates rise 1 point$16,334.23+5.03% a year over 10 years

If rates stay put$17,310.76+5.64% a year: the yield less the fee

If rates fall 1 point$18,351.36+6.26% a year

Lowest point if rates rise−2.8%after 2 years; a loss on paper

Back to what you put in2.5 yearsCounting reinvested income

Rising rates hurt first and help later. The price falls as rates climb, but every bond the fund buys afterwards pays the higher rate. Hold for about the fund’s duration (14.7 years for TLT) and the extra income roughly makes up the price loss. If rates never move, TLT is expected to return about its yield less its fee: 5.50% a year.

History: the starting yield has been the best guess of what bonds return

Yield at the startReturn over the next 10 years

Starting yield and next-decade return
Start	Starting yield	Next 10 years, a year
1961	2.76%	2.89%
1962	2.35%	3.64%
1963	3.85%	3.36%
1964	4.14%	3.56%
1965	4.21%	3.38%
1966	4.65%	3.67%

Across 56 ten-year stretches starting 1961–2016, the 10-year yield on the first day and the yearly return over the next ten years moved together (correlation 0.93, where 1 is lockstep). The gaps are the surprises. After rates peaked in the early 1980s, falling rates let bonds beat their starting yield. When inflation pushed rates up, in the late 1960s and 70s and again in 2022, bonds fell short of it. The 10-year yield averaged 4.97% in Sep 2026.

SHYIEFTLT

Price change by rate change
Rate change	SHY	IEF	TLT
−3.00 pts	+5.54%	+23.06%	+58.11%
−2.50 pts	+4.59%	+18.86%	+46.50%
−2.00 pts	+3.65%	+14.81%	+35.65%
−1.50 pts	+2.72%	+10.90%	+25.58%
−1.00 pts	+1.80%	+7.12%	+16.28%
−0.50 pts	+0.90%	+3.49%	+7.76%

Same direction, very different size. A 1 point rise costs SHY about 1.8%, IEF 6.6% and TLT 13.2%. TLT’s line curves upward (convexity): a 3 point fall gains more than a 3 point rise loses.

The math matches how much they have actually swung

Duration math: 1 point rate moveMeasured: yearly volatility, last 3 years

Predicted and measured risk
Fund	1 point move	Volatility
SHY	1.76%	1.52%
IEF	6.56%	6.57%
TLT	13.20%	13.80%

The measured swings (iShares’ 3-year standard deviation) line up with duration: each fund has been about as volatile as a 1 point rate move a year would make it. Longer bonds are not riskier by accident. They are more sensitive by arithmetic.

After a jump in rates: how long each takes to recover

Rates jump today byThen they stay there. Income is reinvested.

+1.00 pts

SHYIEFTLT

Total return by year
Year	SHY	IEF	TLT
0	−1.8%	−6.6%	−13.2%
1	+4.0%	−0.7%	−7.4%
2	+10.1%	+5.5%	−1.2%
3	+16.5%	+12.1%	+5.4%
4	+23.3%	+19.0%	+12.5%
5	+30.5%	+26.5%	+20.0%

SHY earns back a +1.00 pts jump in 4 months, IEF in 1.2 years and TLT in 2.3 years. Pulling ahead of where each would have been with no change takes longer, because that path kept earning too: SHY 1.8 years, IEF 6.8 years, TLT 14.3 years.

When the Fed “prints” money to buy Treasuries (quantitative easing, QE), it becomes a huge extra buyer. More demand for the same bonds pushes their prices up and yields down, mostly for longer bonds, because the Fed is taking long-term risk off the market. When it lets bonds run off or sells them (quantitative tightening, QT), the public has to absorb more bonds, and yields tend to rise.

Fed bond holdings change by

+$1.00 trillion (buying, QE)

10-year yield moves per $1 trillionResearch on 2008–10 puts it between about 17 and 57 basis points (0.17–0.57 points). 30 is the middle estimate.

30 bp

Inflation worries addPrinting can raise expected inflation, which pushes yields up and works against QE. Applied to every fund equally.

0 bp

SHY price+0.2%Yield −9 bp

IEF price+1.9%Yield −27 bp

TLT price+3.4%Yield −22 bp

SHYIEFTLT

Price change by Fed purchases
$ trillion	SHY	IEF	TLT
-3	−0.47%	−5.37%	−9.19%
-2.5	−0.39%	−4.50%	−7.76%
-2	−0.31%	−3.62%	−6.28%
-1.5	−0.23%	−2.73%	−4.77%
-1	−0.16%	−1.83%	−3.22%
-0.5	−0.08%	−0.92%	−1.63%

In this setting TLT gains most per dollar the Fed spends, even though 30-year yields moved a little less than 10-year yields in 2008–09, because TLT’s duration is twice IEF’s. SHY barely moves: short yields follow the Fed’s interest rate, not its bond buying.

What actually happened in 2008–10

Treasury yield changes around the first round of QE
Window	2-year	10-year	30-year
QE1 announcement days, Nov 2008–Mar 2009 (Gagnon et al.)	−34 bp	−91 bp	—
Same period, every day included (Gagnon et al.)	−19 bp	+50 bp	—
QE1 announcement days, two-day windows (Krishnamurthy & Vissing-Jorgensen)	−39 bp (3-year)	−107 bp	−73 bp

On the days the Fed announced purchases, long yields fell sharply. Over the whole period the 10-year yield still rose 0.50 points, as the economy recovered and inflation fears grew. QE pulls yields down against everything else that moves them; it does not decide the direction alone. The model spreads the effect across maturities in the shape the QE1 days showed (1-year -25 bp, 3-year -39 bp, 5-year -74 bp, 10-year -107 bp, 30-year -73 bp) and assumes selling works like buying in reverse, which is less well measured.

Each bar moves every yield from today’s level (SHY 4.84%, IEF 5.22%, TLT 5.65%) to the rate shown, overnight. Moves this large break the simple duration rule, so these bars reprice every bond in each fund’s maturity range exactly.

SHYIEFTLT

Instant price change
Rates at	SHY	IEF	TLT
−2%	+13.4%	+68.2%	+295.0%
−1%	+11.2%	+56.0%	+221.5%
0%	+9.2%	+44.9%	+163.6%
2%	+5.2%	+25.3%	+81.4%

Lower rates mean a bigger one-time gain, and it grows faster the lower rates go. At 0%, TLT would gain about 164%: a 30-year bond paying 3% is worth far more when new bonds pay nothing. SHY’s bonds come due within three years, so it can only gain a few percent however low rates go.

Then hold forRates stay at the new level the whole time.

10 years

SHYIEFTLT

Yearly return while holding
Rates at	SHY	IEF	TLT
−2%	−0.9%	+3.1%	+12.3%
−1%	−0.1%	+3.3%	+11.1%
0%	+0.7%	+3.6%	+10.0%
2%	+2.4%	+4.2%	+8.1%

The gain happens once; the low rate lasts. At −1%, every year the fund earns −1% less its fee, so the one-time jump is slowly given back. SHY shows it first, because it had little jump to spend. Hold long enough at a negative rate and every fund ends up losing money each year. Negative yields are real: Germany’s 10-year yield stayed below zero from 2019 into early 2022, and Switzerland’s 10-year fell below −1% in August 2019.

Looking ahead: $1 in the S&P 500 against $1 in TLT, on your Fed path

A stock index’s price is its earnings times what buyers pay for each dollar of them (the P/E), so over time stocks return roughly their dividend yield, plus earnings growth, plus or minus any change in the P/E. A bond fund returns its yield, plus or minus price changes when rates move. Map out the Fed moves you expect, and each one moves TLT’s yield by the same amount, just as in section 1.

Bond fund

Years ahead

10 years

S&P 500 earnings growthEarnings grew 6.3% a year from 1950 to Jun 2026 (Shiller).

6.3% a year

P/E in 10 yearsWhat buyers pay per dollar of earnings at the end. The median was 17.7 since 1950 and 22.5 since 1990.

25.2 (today 25.2)

Your rate expectations: the Fed’s rate and TLT’s yield

Both lines start at today’s real levels: the Fed’s rate at 3.88% (the latest saved figure; live data could not be loaded) and TLT’s yield at 5.65%, from iShares. Click the chart to add a move at the size below, as many as you like. Click a dot to turn it into a hike or a cut, or remove it.

Size of a new move (basis points)−25 is a quarter-point cut, +50 a half-point hike.

Fed rate, your pathTLT yield

Rates each time your path changes
From	Fed rate	TLT yield
Sep 2026	3.88%	5.65%
Sep 2036	3.88%	5.65%

No Fed moves yet, so rates stay where they are. Click the chart to add your first.

What $1 returns

S&P 500TLT, your pathTLT, no moves

Total return and yearly average at each year
Year	S&P 500	TLT, your path	TLT, no moves
0	0.0% · 0.0% a year	0.0% · 0.0% a year	0.0% · 0.0% a year
1	+7.5% · +7.5% a year	+5.6% · +5.6% a year	+5.6% · +5.6% a year
2	+15.5% · +7.5% a year	+11.6% · +5.6% a year	+11.6% · +5.6% a year
3	+24.1% · +7.5% a year	+17.9% · +5.6% a year	+17.9% · +5.6% a year
4	+33.4% · +7.5% a year	+24.6% · +5.6% a year	+24.6% · +5.6% a year
5	+43.3% · +7.5% a year	+31.6% · +5.6% a year	+31.6% · +5.6% a year

S&P 500, yearly average+7.46%+105.4% in total over 10 years

TLT, your Fed path+5.64%+73.1% in total; yield ends at 5.65%

TLT, no Fed moves+5.64%+73.1% in total: today’s yield less the fee

Where the stock return comes from1.1% + 6.3% + 0.0%Dividend yield today + earnings growth ± yearly change in the P/E

With these inputs stocks end ahead, by 1.82% a year. Each dot marks one of your Fed moves. A cut lifts TLT’s price that month, then it earns a lower yield from then on; a hike does the reverse. Hover the chart for the total return and the yearly average at any point.

Your Fed path

Add a move in year

No Fed moves: the rate stays where it is. Click the chart or use Add.

Each Fed move moves TLT’s yield by the same amount, as section 1 assumes. In practice long yields often move less than the Fed’s rate, and prices move ahead of the Fed when traders see a move coming, so treat the jumps as the full effect of that change in rates. Stocks and bonds are modelled separately: a Fed move changes the bond fund here, not the stock line. The P/E moves evenly from today’s to your ending value, and dividends stay 28% of earnings, as in Jun 2026. For scale: since 1928, when stocks started a decade in the most expensive fifth of valuations (CAPE above 23; it is 39.8 now), they returned 6.2% a year over it, and 10-year Treasuries 4.5%.

What stocks earn per dollar, next to what bonds pay

S&P 500 earnings ÷ price10-year Treasury yield

Earnings yield and bond yield
Year	Earnings yield	10-year yield
1960	5.34%	2.76%
1961	4.71%	2.35%
1962	5.81%	3.85%
1963	5.51%	4.14%
1964	5.62%	4.21%
1965	5.73%	4.65%

The earnings yield is the P/E turned upside down: what a dollar of stock earns in a year. At a P/E of 25.2 (Jun 2026) that is 3.96%, while the 10-year Treasury pays 4.97% and TLT 5.65%. When the bond line sits above the stock line, buyers are paying for earnings growth that has not happened yet; that happened in the 1980s–90s and again now.

When stocks crashed, did bonds help?

Years the S&P 500 lost more than 10%, and what 10-year Treasuries did
Year	S&P 500	10-year Treasuries	CAPE going in
1931	−43.8%	−2.6%	16.0
2008	−36.5%	+20.1%	25.9
1937	−35.3%	+1.4%	21.1
1974	−25.9%	+2.0%	13.5
1930	−25.1%	+4.5%	21.9
2002	−22.0%	+15.1%	30.5
2022	−18.0%	−17.8%	38.1
1973	−14.3%	+3.7%	18.6
1941	−12.8%	−2.0%	13.9
2001	−11.8%	+5.6%	37.1
1940	−10.7%	+5.4%	16.4
1957	−10.5%	+6.8%	17.1

In 9 of 12 years that stocks fell more than 10%, Treasuries made money, often a lot (2008: +20.1%). They lost money in 1931, 1941 and 2022. In 2022, inflation drove rates up and hit stocks and bonds at once. An expensive market raises the odds that a cushion is needed; it does not guarantee bonds will provide it.

Fund figures: iShares, as of Sep 29, 2026. History: Damodaran and Shiller via this site’s market data archive, read Sep 30, 2026.

Fund figures (duration, convexity, yield to maturity, maturity, fee and three-year volatility) are iShares’ published portfolio characteristics, read from each fund’s page by this site at most twice a day. If they cannot be loaded, the page says so and uses iShares’ figures from September 29, 2026.
Each fund is treated as keeping a constant duration, as its index does by selling bonds that get too short and buying longer ones. It earns its yield less the fee every month, reinvested, and its price moves by −duration × the rate change + ½ × convexity × the rate change squared. Roll-down, the small extra return from bonds ageing along an upward-sloping yield curve, is left out.
Every maturity moves by the same amount in sections 1, 2 and 4. Real yield curves twist: short and long rates often move by different amounts.
Section 4 reprices each fund as a ladder of Treasuries across its maturity range (1–3, 7–10 and 20–30 years) paying the fund’s average coupon, scaled so a small move matches the published duration. The unscaled ladder is within 7% of each fund’s published duration, and within 15% of its convexity allowing for rounding.
Section 3 assumes the Fed’s buying or selling moves the 10-year yield by the set amount per trillion dollars (30 basis points by default, from Gagnon et al.’s range of 30–100 bp for $1.75 trillion), and other maturities in the proportions seen on the 2008–09 QE announcement days. Selling is assumed to work like buying in reverse.
The look-ahead in section 5 prices the S&P 500 as earnings × P/E: earnings grow at the rate you set, the P/E moves evenly (in ratio) from today’s to the one you set, and dividends stay the same share of earnings as in the latest month, reinvested monthly. The Fed’s rate starts at today’s effective rate and you draw every move after that. Each move shifts the fund’s yield by the same amount that month, with the duration-rule price change of section 1. Fed moves do not change the stock line. Returns are shown for $1, with the yearly average.
Historical returns are calendar-year total returns on the S&P 500 and a constant-maturity 10-year Treasury bond from Damodaran. They are not TLT, IEF or SHY, which started trading in 2002. CAPE is computed from Shiller’s monthly data as the real price over the average of 120 months of real earnings.
Taxes, trading costs and the gap between an ETF’s market price and its net asset value are not modelled.

What this tells you

A bond is a loan. You lend money to the US government and it pays you a fixed amount of interest every six months, then hands the money back on a set date. A bond ETF, such as SHY, IEF or TLT, is one share that holds many of these loans at once and buys new ones as the old ones come due.

This page answers five questions about these funds with charts: what happens when interest rates rise or fall, why the three funds behave so differently, what the Federal Reserve's bond buying and selling does, what negative interest rates would mean, and how bonds have fared when stocks were expensive or cheap.

Why bond prices fall when rates rise

Say you own a bond paying 3% a year. Then new bonds start paying 5%. Nobody will pay full price for your 3% bond when a 5% one is on offer, so its price drops until the two are an equally good deal. When rates fall, the opposite happens: your older, higher-paying bond becomes worth more.

Duration measures how much. It is roughly the percent a fund's price changes when rates move by 1 percentage point. TLT's duration is about 15 years, so a 1 point rise costs it about 15%. SHY's is under 2 years, so the same rise costs it under 2%. Longer loans are locked into their old rate for longer, so a rate change matters more to them.

Convexity is a small correction. The price does not move in a straight line: as rates fall, the price rises a little faster, and as they rise, it falls a little slower. It matters most for long funds and for big moves.

Why rising rates are not all bad news

The price drop happens once. But the fund keeps buying new bonds, and after rates rise those new bonds pay more. Over time that extra income makes up the loss. As a rule of thumb, if you hold a fund for about as long as its duration, a rise in rates roughly evens out. Hold longer and you end up ahead of where you would have been.

If rates never change, a bond fund is expected to return about its yield to maturity (what its bonds pay if held to the end) less its fee. The second chart in section 1 shows that since 1960, the 10-year Treasury yield at the start of a decade has been a close guide to what 10-year Treasuries returned over it.

The Federal Reserve buying or selling bonds

When the Fed creates money to buy bonds (quantitative easing, or QE), it becomes a very large extra buyer. More demand for the same bonds raises their prices and lowers their yields. When it shrinks its holdings (quantitative tightening, or QT), other investors have to buy those bonds instead, and they want a higher yield to do it.

Research on the first round of QE, in 2008 and 2009, found long-term yields fell sharply on the days the Fed announced purchases. Over the whole period, though, the 10-year yield rose, as the economy recovered and inflation worries grew. The Fed is one force on rates among several. The model on this page lets you set how strong it is and add an offsetting rise from inflation worries.

Negative rates

A negative rate means lenders pay to lend. It sounds impossible, but German, Swiss and Japanese government bonds all had negative yields in the late 2010s. If US rates fell that far, bond prices would jump, long ones enormously. After that, though, the fund would earn a negative yield every year, slowly giving the gain back. A gain from falling rates happens once. The low income that follows lasts as long as the rates stay low.

Expensive and cheap stock markets

The CAPE ratio is the price of the S&P 500 divided by its average inflation-adjusted earnings over the past ten years. A high number means stocks are expensive compared with what companies have earned. Since 1928, when stocks started a decade cheap, they beat Treasuries every time. When they started expensive, bonds won almost as often as they lost. Bond returns themselves did not depend much on stock prices. What changed was how much stocks had left to give.

The last chart compares the earnings yield (earnings divided by price, the P/E ratio upside down) with the 10-year Treasury yield. It shows how much a dollar in stocks earns next to what a dollar in bonds pays.

Looking ahead: stocks against bonds, with your own Fed moves

The last section lets you set your own expectations. A stock index's price is its companies' earnings times what buyers are willing to pay for each dollar of them, the P/E ratio. So you choose how fast earnings grow and where the P/E ends up. If the P/E falls from today's level, that drag comes off the return every year.

For the bond fund, the rate chart starts from today's real numbers: the Fed's rate and the fund's yield. From there, you map out the Fed moves you expect. Click the chart to add a cut or hike, as many as you like, or click a dot to turn it around or remove it. Each move changes the fund's yield by the same amount, so its price jumps that month, just as in the rate-move chart at the top.

The result compares $1 in each: the total return so far, and the yearly average return it works out to. Hover the chart to read both at any point.

What this does not account for

Rates at every maturity rarely move together. Short and long rates often move by different amounts.
The funds' figures come from iShares' pages, read by this site at most twice a day. If they cannot be loaded, the page says so and uses the figures from September 29, 2026.
In the look-ahead, Fed moves change the bond fund only. In reality they also move stock prices.
Long-term yields often move less than the Fed's rate, and bond prices move ahead of the Fed when traders expect a change. The look-ahead moves the fund's yield by the full amount, on the day.
The historical returns are for a 10-year Treasury bond, not for these three ETFs, which began in 2002.
The Fed model rests on research about one period, 2008–2010. Nobody knows how precisely it applies to future buying or selling, and selling has been studied less than buying.
Taxes, trading costs and gaps between an ETF's market price and the value of its bonds are left out.
Past relationships between stock prices and bond returns are history, not a forecast.