Issue 08: Who Pays for a 5% Treasury? Stocks Hurt Most by Rising Rates
· Members
A December rate hike is priced. Who pays for it is not, and the answer is narrower than most investors think.
Homebuilders are the stocks hurt most by rising interest rates. On an average day this year, they have fallen about 2.7% for each 0.1-point rise in the 10-year Treasury yield that day (say, from 5.12% to 5.22%), about five times as much as the S&P 500.
Homebuilders move about five times as much as the S&P 500 when the 10-year yield rises
Lennar−3.24%
KB Home−3.19%
Toll Brothers−3.15%
D.R. Horton−3.02%
PulteGroup−2.73%
Homebuilders (iShares ETF)−2.68%
NVR−2.02%
Consumer discretionary−1.03%
Russell 2000 small caps−0.97%
Real estate−0.86%
Utilities−0.83%
Regional banks−0.68%
S&P 500−0.53%
Technology−0.52%
Financials−0.43%
Energy1.16%
Average same-day price move for each 0.1-point (10 bp) rise in the 10-year Treasury yield, 2026 trading days through November 13. Sectors are the Select Sector SPDR funds; regional banks are the SPDR S&P Regional Banking ETF.
Source: Yield Theory regression of Yahoo Finance closes on US Treasury 10-year yields
Utilities, real estate and small companies lose about 0.8% to 1%. Banks, the textbook winners from higher rates, have also fallen when yields rose. And the big tech stocks that drive the index have barely reacted to rates since July. They trade on AI.
We measured how every major US stock group has moved with the 10-year yield this year, day by day, because the next move is close: markets put the odds of a quarter-point Fed hike on December 9, to a range of 4% to 4.25%, at about 90%, and the 10-year closed at 5.12% on November 16. Our free explainer covers the decision itself.
Three calls follow from the screen.
Utilities are the most mispriced rate trade. The sector's dividend yield is 2.4 points below the 10-year, the widest gap since at least 2001, yet they still fall when yields rise. Duke Energy and Southern Company are not growing dividends fast enough to justify it.
Homebuilders are priced very differently. Lennar trades at its lowest price to book value since at least 2014, and KB Home sits in the bottom tenth of its range. Toll Brothers trades above 80% of its own history. If yields stay above 5%, the premium builders have further to fall.
Banks are not a hedge. By their own filings, a full one-point rise in rates would add only 1% to 3% to the big banks' net interest income, and their shares have fallen on days when yields rose.
How we measured rate sensitivity
We took every trading day in 2026 through November 13, 214 days in all, and lined up each stock's daily price change against that day's change in the 10-year Treasury yield from the Treasury's daily yield curve. A regression, a standard way of finding the average relationship between two series, gives the typical move for each 0.1-point change in the yield. We call that the stock's rate sensitivity. Prices are Yahoo Finance closes.
Two caveats matter. First, sensitivity is an average, not a law. For homebuilders, the yield explains about a third to two-fifths of daily moves, a strong link for stock data. For tech, it explains almost none. Second, the numbers describe 2026. Since July, the S&P 500's sensitivity has fallen to −0.20% per 0.1 point, and tech's to roughly zero, because AI news has drowned out everything else. For both, the link is now too weak to lean on.
Since July 1
Move per 0.1-point rise in the 10-year
How closely it tracks yields (correlation, −1 to 1)
Homebuilders (iShares US Home Construction)
−2.75%
−0.64
NextEra Energy
−1.34%
−0.55
Utilities
−0.76%
−0.39
Real estate
−0.78%
−0.40
Regional banks
−0.69%
−0.31
S&P 500
−0.20%
−0.11
Technology
+0.09%
0.02
Energy
+1.12%
0.37
What happened the last time the 10-year neared 5%
Before this year, the last time was October 2023; in 2026 the 10-year first closed at 5% on September 15. In 2023 it climbed from 3.75% on July 19 to 4.98% on October 19, then fell back to 3.79% by December 27 as inflation cooled and the Fed signaled it was done raising rates. That round trip, up 1.23 points and down 1.19 points, is the best test of our sensitivities, because each leg is about twice the size of the moves in our scenarios.
Group
Jul 19 to Oct 19, 2023: 10-year up 1.23 points
Oct 19 to Dec 27, 2023: 10-year down 1.19 points
Homebuilders (iShares US Home Construction)
−18.0%
+40.6%
Lennar
−20.0%
+42.7%
Toll Brothers
−15.3%
+49.9%
Regional banks
−13.6%
+30.4%
Real estate
−14.6%
+22.3%
Russell 2000 small caps
−14.3%
+21.5%
Utilities
−11.5%
+7.4%
NextEra Energy
−27.3%
+15.6%
Technology
−7.3%
+15.6%
S&P 500
−6.3%
+11.8%
Energy
+11.2%
−7.0%
Share price changes from Yahoo Finance closes. Yields are the Treasury's 10-year par yields, via FRED.
Three lessons carry over to today.
The broad order holds. Builders moved most, the S&P 500 least among the groups that fell, and energy moved the other way. Banks, real estate, small caps and utilities sat in between, as they do in 2026. Single stocks can stray far from their group, though: NextEra fell further than the builders.
Builders fall less than the math says, and rebound more. Applying the 2026 sensitivity to a 1.23-point rise implies a 33% fall for the homebuilder fund. It fell 18%. On the way back down it rose 41%, more than the 32% the same sensitivity implies. Once builders get cheap, buyers step in, and when rates turn, the rally is fast. We read the bear case prices below as more severe than likely, and we see the rebound as the bigger opportunity.
Defensive groups fell more than the math says. The straight line implied falls of 10.2% for utilities, 10.6% for real estate and 8.4% for regional banks. They fell 11.5%, 14.6% and 13.6%. Utilities then regained only 7.4% in the two months after yields peaked, and NextEra lost 27% and regained 16%. For the banks, that was months after Silicon Valley Bank failed, when deposit costs and bond losses dominated, but the direction matches 2026.
Our view by group
Group
Move per 0.1-point rise in the 10-year
Valuation against its own history
Our view
Homebuilders
−2.0% to −3.2%
Lennar and KB Home near decade lows; Toll and PulteGroup above their medians
Avoid the premium names; Lennar once its margins stop falling
Utilities
−0.8%
Dividend yield 2.4 points under the 10-year, the widest since 2001
Underweight Duke and Southern
Real estate
−0.9%
Yield 1.9 points under the 10-year, against an average of 0.4 points over since 2016
Neutral; refinancing costs are manageable
Banks
−0.2% to −0.7%
JPMorgan at 3.0 times tangible book
Not a hedge
S&P 500
−0.5%
Forward P/E 17.8, below its 10-year average
Earnings carry it, but the cushion over bonds is thin
Energy
+1.2%
Valuation not covered here
The one group that rises with yields
Can the S&P 500 live with a 5% Treasury?
For now, yes, because earnings are doing the work. FactSet expects S&P 500 earnings for the third quarter to be up 31.9% from a year earlier, the third straight quarter above 25%. Its consensus for the next 12 months was $415.82 a share on November 5.
At the November 16 close of 7,416.59, that is a forward price-to-earnings ratio (P/E) of 17.8. Turn it upside down and you get the forward earnings yield: 5.61%. That is what the index is expected to earn for each dollar invested, before any growth.
The comparison with bonds is where it gets uncomfortable. The 10-year Treasury paid 5.12% at Monday's close, with no risk of default. Stocks are expected to earn only 0.49 points more.
Measure
Now
10-year average
5-year average
Forward P/E (FactSet)
17.8
19.1
19.8
Forward earnings yield
5.61%
5.24%
5.05%
10-year Treasury yield
5.12%
2.94%
3.91%
Gap: earnings yield minus Treasury
0.49 points
2.30 points
1.14 points
The averages for the P/E come from FactSet; the Treasury averages are our calculation from FRED daily data for the same windows.
So the index looks cheap against its own history and expensive against bonds. Investors are paying less for each dollar of expected profit than they used to, but those profits are expected to grow fast, and the safe alternative now pays far more than at any time in the past 10 years.
How much does the yield matter? The table below holds forward earnings at $415.82 and asks what the S&P 500 would be worth at different 10-year yields and different gaps between the earnings yield and the bond yield. It is a sensitivity, not a forecast: it ignores earnings growth over the year, which on FactSet's numbers is large.
10-year yield
Gap stays at 0.49 points
Gap widens to 1.0 point
Gap widens to 1.5 points
4.50%
8,339 (P/E 20.1)
7,560 (18.2)
6,930 (16.7)
5.12%
7,417 (17.8)
6,794 (16.3)
6,281 (15.1)
5.75%
6,667 (16.0)
6,160 (14.8)
5,735 (13.8)
Read across the middle row. With yields where they are, the index is priced for the gap to stay near its thinnest in years. If investors demanded even a one-point cushion, the index would need to fall about 8%, or forward earnings would need to rise about 9% to hold today's price.
Our daily measure tells a gentler story: −0.53% for each 0.1-point rise across 2026, so a 0.6-point rise in yields would cost about 3%. The difference is time. Day to day, AI news swamps rates. Over months, valuations have to answer to the bond market.
One more detail shows where the index's rate risk hides. The equal-weight S&P 500, which gives every company the same weight, has been more sensitive than the normal index: −0.73% per 0.1 point this year against −0.53%. The giant AI companies at the top of the index mask how much the typical stock cares about yields.
Homebuilders: the most exposed group, priced two different ways
Why builders move so much
Builders sell houses to people who borrow to buy them. Freddie Mac's average 30-year mortgage rate was 7.11% in the week to November 12, against 6.24% a year earlier. On a $400,000 loan, that adds about $230 to the monthly payment.
Builders have been paying to close that gap. They cut prices, cover closing costs and pay to buy down buyers' mortgage rates. Every dollar spent that way comes out of their profit margin. Their filings for the latest quarter show the damage.
The outlook is worse than the last quarter. D.R. Horton guided to a home sales gross margin of 18.0% to 18.5% for the quarter to December, down from 20.7% in the quarter to June, the latest quarter for which it has disclosed the figure, and said it expects "incentives to remain elevated in fiscal 2027." Lennar guided to 15.5% to 16.0% and cut its delivery target for the year to about 80,000 to 81,000 homes. Its 10-Q put the cause plainly: "near-term interest rate relief appears unlikely." NVR, which sells mostly to buyers who order before it builds, said "demand for new homes weakened further" in the third quarter.
How far mortgage rates can fall without the Treasury
The Fed does not set mortgage rates. The 30-year rate tracks the 10-year Treasury plus a spread, which pays lenders and bond investors for, among other things, the risk that borrowers refinance early. In the week to November 12, that spread was 2.01 points: a 7.11% mortgage rate against a 10-year yield of 5.10% that day. From 2010 to 2019 it averaged 1.69 points, and in July 2023 it reached 3.20.
So there is a little room. If the spread went back to its pre-pandemic average with the 10-year unchanged, the 30-year rate would be about 6.8%, still more than half a point above a year ago. For the builders, the 10-year yield remains the number that matters. Spreads are from Freddie Mac's weekly survey history and Treasury yields, our calculation.
What the shares already price in
The group splits on valuation. The usual way to value a builder is price to book value (P/B): the share price divided by the company's net assets per share, which for a builder are mostly land and houses. We computed each builder's P/B at every quarter end since early 2014 from its SEC filings and Yahoo Finance prices.
Builder
Price, Nov 16
P/E, last 12 months
P/B now
Median P/B since 2014
Rank in its own history
Return on equity, last 12 months
Lennar
$83.83
15.9
0.93
1.31
Lowest
5.8%
KB Home
$47.40
13.4
0.76
1.04
8th percentile
5.7%
NVR
$5,795
16.0
4.46
5.23
16th percentile
27.9%
D.R. Horton
$144.76
14.3
1.70
1.79
34th percentile
11.9%
PulteGroup
$127.01
13.6
1.79
1.65
67th percentile
13.2%
Toll Brothers
$147.19
11.8
1.59
1.32
80th percentile
13.5%
Book values are at each builder's latest quarter end; Lennar's and KB Home's are at August 31, Toll's at July 31. Return on equity (ROE) is last-12-months diluted earnings per share divided by book value per share. Earnings per share for the last 12 months: D.R. Horton $10.11 (fiscal year to September), Lennar $5.28, PulteGroup $9.36, NVR $362.71, Toll $12.46 and KB Home $3.55, summed from quarterly filings. For Lennar, Toll and KB Home, the fiscal 2025 fourth quarter inside each 12-month window is derived as the full year minus the first nine months.
A low P/B is not automatically cheap. A builder that earns a low return on its equity deserves a low P/B. So the useful question is what return each price assumes.
If investors want 10.12% a year from a builder (the 5.12% Treasury yield plus a 5-point premium for owning a cyclical stock), and book value grows about 3% a year, then a P/B of 1.0 means the market expects the builder to earn about 10.12% on its equity. Each extra turn of P/B means it expects more. Formally, the ROE a price implies is 3% + P/B × (10.12% − 3%). We compare that with what each builder actually earned before the pandemic, from 2016 to 2019, a stretch of steady housing demand and mortgage rates of about 4%.
Builder
ROE the price implies
ROE in the last 12 months
Average ROE, 2016 to 2019
What the price assumes
Lennar
9.6%
5.8%
About 12.6%
Less than its pre-pandemic normal
KB Home
8.4%
5.7%
8.3%
Its pre-pandemic normal
D.R. Horton
15.1%
11.9%
15.5%
Its pre-pandemic normal
PulteGroup
15.7%
13.2%
16.0%
Its pre-pandemic normal
NVR
34.8%
27.9%
35.8%
Its pre-pandemic normal
Toll Brothers
14.3%
13.5%
11.9%
More than its pre-pandemic normal
Pre-pandemic ROE is the fiscal-year diluted EPS divided by the average of opening and closing book value per share, from SEC data. Lennar's is approximate because its 2018 merger with CalAtlantic changed its share count mid-year.
Three readings follow.
Most of the group is priced for a return to normal profits. D.R. Horton, PulteGroup and NVR are each priced for a return on equity close to what they earned in 2016 to 2019. They are earning about 3 to 8 points less today, and their own guidance says margins fall further next quarter. That return to normal needs mortgage rates to come down. A 10-year yield above 5% pushes it out.
Toll Brothers is priced above its own normal. Toll sells larger, more expensive homes, and its ROE has held up better than the group's. Even so, its price implies a 14.3% ROE against 11.9% before the pandemic, and its P/B sits higher than 80% of its quarter ends since 2014. Its shares are also among the most rate-sensitive in the group at −3.15% per 0.1 point. Its trailing P/E of 11.8 is the lowest in the group, but builders often look cheapest on earnings near the top of a cycle, which is why we anchor on book value. A premium valuation on book with a high rate sensitivity makes Toll the most exposed name in the sector if yields stay high.
Lennar is priced below its own normal. At 0.93 times book, Lennar trades lower than at any quarter end since 2014 and below the ROE it earned before the pandemic. It is also the most rate-sensitive stock we measured. That makes it the cleanest way to bet on falling rates. On earnings it is not cheap, at 15.9 times the last 12 months, because its profits have fallen further than its peers'. The case rests on book value and on margins recovering. But its margins are still falling, it has cut its delivery target, and its fiscal fourth quarter, reported in December, will show whether 15.5% is the floor. We would wait for that before buying.
The sensitivity to our cost of equity assumption is modest for the cheap names and large for the expensive ones. Raise the required return by half a point, to 10.62%, and the ROE NVR's price implies rises by 2.2 points, Lennar's by only 0.5 points. That is another way of saying that high-P/B builders carry more valuation risk from higher rates.
Three rate paths for the builders
The table applies each stock's 2026 rate sensitivity to the three yield paths we set out in the scenario section below. It is the rate effect alone, holding everything else equal, and then shows what P/B the bear case price would mean.
Builder
Bull: 10-year to 4.50%
Base: 10-year at 5.12%
Bear: 10-year to 5.75%
Bear, if the fall is half as large, as in 2023
P/B at the bear price
Lowest P/B since 2014
Lennar
$101 (+20%)
$84
$67 (−20%)
$75 (−10%)
0.74
0.93 (now)
KB Home
$57 (+20%)
$47
$38 (−20%)
$43 (−10%)
0.61
0.63
D.R. Horton
$172 (+19%)
$145
$117 (−19%)
$131 (−10%)
1.38
1.18
PulteGroup
$149 (+17%)
$127
$105 (−17%)
$116 (−9%)
1.48
1.05
NVR
$6,521 (+13%)
$5,795
$5,058 (−13%)
$5,426 (−6%)
3.90
3.82
Toll Brothers
$176 (+20%)
$147
$118 (−20%)
$133 (−10%)
1.27
0.66
The bear case takes Lennar and KB Home to valuations below anything in the last 12 years, which suggests their sensitivity would probably weaken as they got cheaper. That is what happened in 2023, when the builders fell about half as much as their sensitivity implied. It leaves Toll Brothers at about its long-run median P/B, with room to fall further. That asymmetry is the core of our homebuilder view: if you want rate exposure, take it where the valuation already assumes bad news.
Utilities: priced like growth stocks, trading like bonds
Utilities are supposed to be the safe, boring corner of the market: regulated businesses that pay steady dividends. Investors often treat them as a substitute for bonds. That is why they fall when bond yields rise: −0.83% per 0.1-point rise in 2026, and more for NextEra.
The problem is that utilities now pay much less than bonds. The Utilities Select Sector SPDR fund yields 2.74% on its last 12 months of dividends. The 10-year Treasury yielded 5.12% at Monday's close. The gap, −2.38 points, is wider than at any month end in our data going back to January 2001.
Utility dividends now pay 2.4 points less than the 10-year Treasury, the widest gap since at least 2001
0123456%200120102019NowGap of −2.38 points2.74%Utilities fund dividend yield5.12%10-year Treasury yield20013.14%5.07%20043.14%4.24%20072.59%4.04%20104.06%3.30%20133.86%3.04%20163.42%2.45%20192.95%1.92%20222.92%3.88%20252.71%4.18%Now2.74%5.12%
Year-end values, and November 16, 2026 for now. The dividend yield is the Utilities Select Sector SPDR fund's last 12 months of dividends divided by its price.Source: Yield Theory calculation from Yahoo Finance prices and dividends, and FRED 10-year yields
There have been negative gaps before. In the mid-2000s utilities yielded 1.1 to 1.7 points less than Treasuries and did fine, because their earnings and dividends grew. So the question is whether today's utilities are growing fast enough.
A simple test helps. An investor who wants a 3-point premium over the 10-year for owning a utility needs a total return of 8.12% a year. We use 3 points here, against 5 for the builders, because regulated utility earnings are far steadier. With a dividend yield of about 3%, the dividend has to grow about 5% a year, forever, to deliver that. With a 2-point premium, the hurdle falls to about 3.8% to 3.9% a year for Duke and Southern, still above the dividend growth they are delivering.
Dividends are the latest quarterly rate times four, from NextEra's, Duke's and Southern's third-quarter 10-Qs. NextEra's P/E uses its adjusted EPS of $4.02 for the last 12 months, because its reported earnings swing with derivative gains; Duke's uses $6.84 and Southern's $4.37 of reported diluted EPS.
NextEra clears the bar: it raised its dividend 10% this year. Duke and Southern do not. Their dividends grew 2% to 3%, roughly half what their prices assume.
Their own history says the same. On the last 12 months of dividends, Duke yields 3.29%, lower than at 98% of month ends since 2015, when its median was 4.09%. Southern yields 3.17%, lower than at 89% of month ends, against a median of 4.11%. Investors are paying more for each dollar of these two companies' dividends than at almost any time in a decade, while the 10-year Treasury pays more than 5%, a level it never closed at from 2008 until this year. These trailing yields run slightly below the table's, which use the latest dividend rate. NextEra is the exception: its 2.86% sits above its median of 2.54%. These yields are our calculation from Yahoo Finance monthly closes and dividends.
To be fair to the bulls, utilities are no longer just bond substitutes. Data centers need power, and the companies are spending heavily to supply it. Duke plans $19.5 billion of capital spending in 2027 and $21.2 billion in 2028, and NextEra has committed to about $29.4 billion of projects in 2027. If those investments earn their allowed returns, earnings growth can outpace dividends.
But that spending has to be paid for, and that is where rates bite again. Duke's annual report says it plans to issue $10 billion of new shares from 2027 to 2030, about 10% of its current market value. Its interest expense over the last 12 months was $3.84 billion, up from $3.63 billion for 2025, and it has $3.6 billion of debt maturing in 2027 that will be refinanced at today's higher rates.
Utility
Total debt, Sep 30
Debt due in 2027
Interest expense, last 12 months
NextEra Energy
$106.0 billion
$10.3 billion
About $4.4 billion (our estimate, excluding derivative marks)
Duke Energy
$90.4 billion
$3.6 billion
$3.84 billion
Southern Company
$73.1 billion
$3.4 billion
$3.31 billion
Debt due in 2027 is as of December 31, 2025, from each company's annual report; the 10-Qs don't update the schedule. NextEra's reported interest expense is distorted by gains on interest rate hedges, so we estimate the underlying figure from its derivative footnotes.
On P/E, they are not extreme: Duke trades at 19.1 times against a median of about 21 since 2015, and Southern at 21.7 against about 21, on reported year-end earnings (our calculation from SEC data). The extreme is against bonds. Investors are paying a normal multiple for slow-growing dividends when Treasuries pay 5%.
Our call: underweight Duke Energy and Southern Company while the 10-year stays above 5%. NextEra's dividend growth earns its yield, but it is the most rate-sensitive utility we measured (−1.34% per 0.1 point since July) and carries the uncertainty of its pending all-stock merger with Dominion Energy, so we would not add to it either.
Real estate: the refinancing bill is smaller than the valuation risk
Real estate investment trusts (REITs) own property and pay out most of their income as dividends. Like utilities, they trade partly as bond substitutes, and the Real Estate Select Sector SPDR fund's yield of 3.26% is now 1.86 points below the 10-year. Its month-end average since 2016 was 0.44 points above.
The fear with REITs is refinancing: debt taken out at 2% to 3% in earlier years has to be rolled over at today's rates. We checked how big that bill is for three large REITs.
Funds from operations (FFO) is the standard REIT measure of cash earnings. We use each company's own version: adjusted FFO of $4.42 a share for Realty Income, core FFO of $6.17 for Prologis and adjusted FFO of $11.05 for American Tower over the last 12 months, from their third-quarter10-Qs and releases. The 6.12% refinancing rate is our assumption: the 10-year Treasury plus one point. American Tower's 2027 maturities include euro bonds at a range of rates, so we did not compute a single figure.
The bill is real but small: 1.4% to 2.3% of a year's cash flow per share for Prologis and Realty Income. Refinancing happens slowly, a slice each year, and these companies have weighted average debt terms of five to eight years.
The bigger risk is valuation. Property values move with interest rates, and the sector's sensitivity of −0.86% per 0.1 point reflects that. Realty Income, at 13.2 times cash flow with a 5.56% dividend yield that pays out about 73% of cash flow, already offers a yield above the 10-year. Prologis, at 20.7 times, is priced for continued rent growth from warehouses.
Against their own history, all three yield more than usual. On the last 12 months of dividends:
REIT
Dividend yield, last 12 months
Median since 2015
Share of month ends with a lower yield
Realty Income
5.55%
4.61%
85%
Prologis
3.31%
2.80%
80%
American Tower
3.71%
1.94%
90%
Our calculation from Yahoo Finance monthly closes and dividends, January 2015 to November 16, 2026.
Part of that is the rate regime. For most of this history the 10-year yielded less than 4%, so higher REIT yields partly reflect higher rates rather than bargains, and as a group they still yield less than bonds: the sector fund pays 1.86 points under the 10-year. Realty Income, at 5.56%, is the exception. We cover three of the largest, in net-lease property, warehouses and cell towers; office and apartment REITs carry different risks and are not covered here. Our call: neutral on the group, with a preference for Realty Income's covered yield, near the top of its own range, over Prologis's multiple while yields stay high.
Banks are not the hedge you think
The textbook says banks gain when rates rise, because they earn more on loans faster than they pay more on deposits. Their own filings say the gain is now small.
Every large bank discloses how its net interest income (NII), the difference between what it earns on loans and securities and what it pays on deposits, would change over 12 months if rates jumped instantly. Here is what they reported at September 30.
Sensitivities are from JPMorgan's, Bank of America's, Wells Fargo's and Citigroup's third-quarter 10-Qs. JPMorgan's figure is its earnings-at-risk measure, which also includes some rate-sensitive fees. Yearly NII is third-quarter NII times four. The EPS effect assumes a 25% tax rate. Tangible book excludes goodwill and other intangible assets.
A December hike is a quarter of a point, not a full point. On these numbers it adds about 1.5% or less to any of these banks' earnings. That is not nothing, but it is easily swamped by credit losses, trading revenue and deposit costs, which is why bank shares have fallen slightly on days when yields rose this year.
Two details sharpen the point. Almost all of Citigroup's sensitivity comes from other currencies: a one-point rise in US dollar rates adds just $37 million to its NII by its own table, or about 2 cents a share after tax. The rest of its +$0.64 comes from rates in other currencies, which a Fed hike does not move. And Charles Schwab, whose earnings depend on client cash, warns that "during periods of rapidly rising interest rates, clients tend to reallocate cash out of sweep products" into higher-paying funds. Its simulated gain from a one-point rise is 3.0% of net interest revenue.
If you want a bank for higher short-term rates, Wells Fargo shows the most leverage in its disclosures, including +$1.3 billion from a one-point rise in short-term rates alone, at 12.0 times earnings and 1.9 times tangible book. JPMorgan, at 3.0 times tangible book, already carries a premium valuation. Our call: don't buy banks as a hedge against higher yields. Own them, if at all, for credit quality and capital returns.
Why big tech shrugs off rates, for now
In theory, technology stocks should be among the most rate-sensitive, because their value rests on profits far in the future, and a higher discount rate shrinks distant profits most. Yet in 2026, tech has barely tracked yields: −0.52% per 0.1 point for the year and roughly zero since July. AI news has dominated. The biggest moves of the month had nothing to do with rates: the Nasdaq fell 5.52% on November 5 as AI trades unwound, and Nvidia lost 22.0% from its November 4 close to its November 16 close, a stretch in which founder Jensen Huang stepped down as CEO.
That can change. When AI stocks are falling for their own reasons, a high bond yield removes the safety net of a valuation that looks justified against bonds. Fed Chair Kevin Warsh sketched that risk as one possible scenario on November 16: after heavy investment and stock gains, "expected returns could turn down and equity valuations could fall." Our AI capex tracker follows the spending that has to keep paying off for that not to happen.
Three paths for the 10-year yield
Our scenarios cover the next six to twelve months. In the base case, futures already price the hikes, so the 10-year has little reason to move from 5.12%. The probabilities are our judgment. The stock moves are the rate effect alone, using each group's 2026 sensitivity, plus 2023-adjusted rows for three groups; they do not include changes in earnings.
Bull: one and done
Base: two more hikes
Bear: inflation wins
Our probability
25%
50%
25%
What happens
Core inflation slows to 0.2% a month by early 2027 and oil eases; the Fed hikes in December, then holds
The Fed hikes in December and once more by March, to 4.25% to 4.5%, in line with futures
Oil-driven inflation keeps rising; the Fed hikes faster, toward 5%
10-year Treasury
4.50%, last seen in July
About 5.12%
5.75%, last seen in November 2000
S&P 500, daily sensitivity
+3.3%
0%
−3.3%
S&P 500, valuation table
About 8,340 if the gap holds
About 7,417
About 6,670 if the gap holds
Homebuilders (iShares ETF)
+17%
0%
−17%
Utilities
+5%
0%
−5%
Real estate
+5%
0%
−5%
Regional banks
+4%
0%
−4%
Energy
−7%
0%
+7%
Homebuilders, if 2023 repeats
+21%
0%
−9%
Utilities, if 2023 repeats
+4%
0%
−6%
Real estate, if 2023 repeats
+12%
0%
−7%
The 2023 rows scale each straight-line move by how far the group actually moved in the matching leg of the 2023 round trip. One episode is thin evidence, but it points the same way as valuation.
How much a move in the 10-year has meant for each group in 2026
Pick a group and where the 10-year yield goes
Rate effect on the group+16.6%0.0%−16.9%+5.1%0.0%−5.2%+5.3%0.0%−5.4%+4.2%0.0%−4.3%+3.3%0.0%−3.3%−7.2%0.0%+7.3%
−16.9%+16.6%
Each group's 2026 move per 0.1-point change in the 10-year yield, times the change from 5.12%. Ignores earnings changes and assumes the 2026 relationship holds. Illustrative.
Because the rate paths are roughly symmetric, the probability-weighted straight-line effect is close to zero for every group. That is by construction, not a finding. The 2023 rows are where the case for acting comes from: on that evidence, builders have more to gain than to lose from here, and utilities have more to lose than to gain. That asymmetry, plus valuation, is why we prefer the cheapest builders to Duke and Southern.
Two cautions. First, the S&P 500 rows give very different answers, a 3% move on one method and a 10% move on the other. Over a few days, the daily sensitivity is the better guide, though since July even it has overstated the link. Over a year, the valuation table matters more, because a 0.49-point cushion over bonds cannot shrink much further. Second, energy is the only group that rises with yields this year, most likely because oil has been pushing up both energy profits and inflation. It is the natural hedge for the bear case.
What would change our view
October PCE inflation (November 25) and November jobs (December 4). A soft pair would make the bull case more likely, and Lennar the best performer in the group.
The December dot plot. If the median official pencils in two or more hikes for 2027, the bear case gains weight and utilities and premium builders are most exposed.
Lennar's fiscal fourth-quarter results in December. A gross margin above its 15.5% to 16.0% guidance, with orders holding up, would suggest the margin floor is in, and would make Lennar a buy even without a rate cut.
Duke Energy's 2027 plan. Faster dividend growth, or a smaller share issuance, would weaken our utilities call.
Oil. A peace deal in the Iran war that sends oil sharply lower would cool inflation fast, and would hurt the one group, energy, that has been rising with yields.
Risks to these calls
The biggest risk is the method itself. Sensitivities measured over 214 days can change, and they changed this year: the S&P 500's fell by more than half after July. A 0.6-point move in yields is much larger than any single day in the sample, and relationships can bend at extremes. The 2023 round trip shows they do: builders fell less and rebounded more than a straight line predicts, and utilities did the reverse.
The utilities call could be wrong if AI power demand lifts earnings faster than dividends, as the companies argue. The homebuilder call could be wrong if Washington acts on housing affordability, or if mortgage spreads narrow on their own and mortgage rates fall even with Treasury yields high. And the banks call could be wrong if the yield curve steepens sharply, since long-term rate rises help some banks more than the parallel shocks in their disclosures suggest.
Our read
A 5% Treasury is not a problem for the whole market. It is a problem for a few groups, and the market is pricing them unevenly.
The homebuilders carry the most rate risk by far, and most of them are priced for a return to pre-pandemic profitability that needs lower mortgage rates. Toll Brothers is the most exposed: premium valuation, high sensitivity. Lennar is the opposite, already priced below its own normal, and the stock to buy once its margins stop falling.
Utilities are the quiet mispricing. Investors are paying normal multiples for 2% to 3% dividend growth at Duke and Southern while Treasuries pay 5.12%, and both companies need to borrow and issue shares to fund their building plans. We would hold less of them until yields fall or dividend growth picks up.
Banks don't protect you from higher yields, and the index's protection is thinner than its P/E suggests. If you want a hedge for the bear case, energy has been the only group that rises with yields this year.
How we built this
Rate sensitivities are least-squares regressions of each stock's or fund's daily percentage price change on the same day's change in the 10-year Treasury par yield, over the 214 trading days of 2026 through November 13 for which both exist, and separately from July 1. Prices are Yahoo Finance closes adjusted for splits and dividends; yields are from the US Treasury, cross-checked against FRED. Prices in the tables are closes on November 16, 2026, and the current 10-year yield is that day's close of 5.12%. Treasury's yields are based on quotes taken at about 3:30 p.m. ET, half an hour before stocks close, which slightly weakens the measured daily link. The 2023 test uses closing prices on July 19, October 19 and December 27, 2023, chosen as the low, high and next low of the 10-year yield. Mortgage spreads are Freddie Mac's weekly 30-year fixed rate minus the 10-year Treasury yield on the same day.
Company figures come from SEC filings and earnings releases for the latest reported quarter, as dated in each table. Book value per share is shareholders' equity divided by shares outstanding where the company does not report it; for Lennar's history we use diluted weighted shares. Historical P/B uses each quarter-end's book value and the closing price on that date, from 2014 to 2026. The implied ROE uses a 10.12% cost of equity (the 5.12% 10-year close on November 16 plus a 5-point premium, our assumption) and 3% growth in book value. Company dividend yields in the tables use the latest dividend rate; fund yields and the history comparisons use the last 12 months of dividends. Scenario probabilities are our judgment.
Go deeper
Will the Fed raise rates in December?: the odds, the inflation and jobs data, what each official has said, and what a hike means for mortgages and savings.
NVIDIA has found a way to turn knowledge of its own roadmap into a lower cost of capital for its customers. The result could move AI infrastructure from venture equity into private credit, insurance balance sheets and securitized debt.
Yield Theory is independent stock research for people investing their own money. Here is what's free, what members get, and three pieces to read first.
The thesis, the numbers behind it, and what would break it. Full access is $39 a month.