Vann Equity Management

Market Watch

When a 19-Year High in Bond Yields Meets a Record August

Vann Equity Management 16 min read

The S&P 500 fell 1.4% last week and its 3-week winning streak ended. The Federal Reserve did not meet, did not vote, and did not change its policy rate. Long-term borrowing costs climbed to their highest level since 2007 regardless.

This article is educational only and does not recommend buying, selling, or avoiding any security.

Those facts sit together more comfortably than they look. The Fed controls one interest rate. The bond market sets the rest, and last week it repriced them without waiting for anyone’s permission. That distinction is the whole story, and it’s worth understanding, because it governs a lot more than 5 trading sessions.

Where the year stands

Last week was a down week inside a year that has gone well, and both halves of that sentence matter.

S&P 500 month-end closing level, December 2025 through August 21, 2026 The S&P 500 closed 2025 at 6,845.50, dipped to 6,528.52 at the end of March, then rose to 7,580.06 by the end of May. June and July were essentially flat at 7,499.36 and 7,489.72. The index set a record close of 7,798.99 on August 13 and finished August 21 at 7,674.37. 6,500 7,000 7,500 Record close, August 13: 7,798.99 6,528.52 7,674.37 Dec Jan Feb Mar Apr May Jun Jul Aug Month-end closing level; August shown through the close on August 21. Index level, excludes dividends.

The shape of 2026 has been 3 distinct stretches. A weak first quarter that bottomed at 6,528.52 on March 31. A sharp recovery through April and May that added better than 16% in 2 months. Then a summer that went sideways: June closed at 7,499.36, July closed at 7,489.72, a difference of about a tenth of a percent.

August broke the stall. The index set a record close of 7,798.99 on August 13, and even after last week’s decline it finished Friday at 7,674.37. That leaves August up roughly 2.5% with a week to go, and the year up about 12.1% from the 2025 close of 6,845.50.

One clarification on those figures, because it changes what they mean. They track the change in the index level and leave out dividends. In plain terms, an investor who held these companies and collected their dividends would have done somewhat better than the number shown. Index levels are the standard way returns get quoted, but they aren’t the whole of what an investor earns.

So last week’s 1.4% decline took the index about 1.6% below its record. That’s a pullback inside an advance, not a reversal of one.

What happened last week

The damage was uneven, and the pattern is the tell.

Index Week ended Aug 21 Friday close
Dow Jones Industrial Average -0.85% 53,277.01
S&P 500 -1.43% 7,674.37
Nasdaq Composite -2.05% 26,180.46

Small companies landed in between, with the Russell 2000 down 1.65% on the week.

Friday itself was a good day. The Dow gained 517.60 points, or 0.98%, and the S&P 500 added 33.14 points, or 0.43%, after the Treasury Department said it would double the maximum size of its debt buybacks, a step that eased some pressure on bond prices. The week still finished in the red, and the Nasdaq took more than twice the Dow’s loss.

Tuesday was the worst of it. The S&P 500 fell 0.69%, the Nasdaq dropped 1.33%, and the Philadelphia Semiconductor Index tumbled 5%. Data storage names took the sharpest hits, with Sandisk down 9% and Western Digital down 7.4%. Information technology was the biggest drag on the S&P 500 of any of its 11 sectors.

None of that came from an earnings miss or a company-specific problem. It came from the bond market.

The rate the Fed doesn’t set

Here’s the mechanism, and it’s the part most headlines skip.

The Federal Reserve sets one rate: the federal funds rate, which is what banks charge each other for overnight loans. It’s held at 3.50% to 3.75%, unchanged at the July 29 meeting. That’s the short end of the curve, and the Fed controls it directly.

Investors in the open market price everything longer. When you buy a 30-year Treasury bond, you’re lending the government money for 30 years, and you decide what return you need to make that worthwhile. Nobody sets that number for you. If enough lenders decide they need more compensation, the price of existing bonds falls and their yield rises, and that happens whether the Fed likes it or not.

Glass office towers in a financial district photographed from street level against a clear blue sky.
Illustrative. The Federal Reserve sets one overnight rate. Investors trading in the open market set every longer one.

That’s what happened. The 30-year Treasury yield pushed above 5.3% last week, its highest level since 2007, a 19-year high. The 10-year moved up near 4.7%. Both climbed while the Fed’s policy rate sat still.

The extra return investors demand for lending long instead of short has a name: the term premium. Put simply, it’s the price of uncertainty. Lend for 30 years and a lot can go wrong that wouldn’t in 30 days: inflation could erode what you get back, the borrower could issue far more debt than expected, the political situation could change. When those worries grow, the term premium widens, and long-term rates rise on their own.

This wasn’t a US-only event, which is the strongest evidence it wasn’t about the Fed. Japan’s 10-year borrowing cost reached a 3-decade high slightly under 3%. Long-dated yields in Germany, France and the United Kingdom hit multi-year highs in the same stretch.

Competing for the same pool of savings

The deeper explanation is a supply problem, and it has 3 sources pulling on the same pool of savings at once.

Government borrowing. US federal debt is approaching $40 trillion, and the July deficit alone came to $432 billion. Every dollar of that deficit gets financed by selling Treasury bonds. More bonds for sale, with no matching increase in buyers, means sellers have to offer a higher yield to clear the market.

A large neoclassical government building in bright daylight, photographed from a low angle against a clear sky.
Illustrative. The Treasury finances every dollar of the federal deficit by selling bonds, adding to the supply competing for buyers.

Corporate borrowing for AI. Technology companies are financing data centers and computing capacity on a scale that requires the bond market, not cash on hand alone. That’s capital spending, or capex, meaning money laid out for long-lived physical assets rather than day-to-day operations. Those bonds compete for the same investors who buy Treasuries.

Rows of industrial cooling units on a data center rooftop under bright daylight.
Illustrative. Cooling capacity at this scale is part of what technology companies are financing through the bond market, placing their borrowing in competition with the Treasury's.

Inflation that hasn’t finished falling. Inflation remains above the Fed’s 2% target. A lender being repaid in 2056 needs compensation for whatever inflation does between now and then, and recent readings haven’t made that easier to estimate. Oil prices climbing on the unresolved US-Iran situation added to it.

A quiet suburban residential street with mature trees on a bright clear morning.
Illustrative. Every household eventually feels borrowing costs, whether through a mortgage rate, an auto loan, or the price of everyday spending.

Put those together and you get a straightforward result: more borrowing, competing borrowers, and less confidence about the value of future dollars. Yields go up.

Why technology took the worst of it

This is where the bond market reaches into the stock market, and it explains the table above.

A stock’s value is the company’s expected future profits, converted into what they’re worth today. That conversion uses an interest rate. In plain terms, a dollar you’ll receive in 10 years is worth less than a dollar today, and the higher the interest rate, the less that future dollar is worth right now.

Fast-growing technology companies have most of their expected profits far out in the future. Established industrial, healthcare and consumer businesses have more of theirs in the present. So when long-term rates rise, the far-off profits lose more value than the near-term ones. Same rate move, unequal effect, and that’s the arithmetic behind a Nasdaq that fell 2.05% while the Dow fell 0.85%.

There’s a general portfolio-construction point buried in that. Two holdings can look diversified by industry label and still respond to the same underlying variable in the same direction. Last week, “technology” and “long-duration interest rate exposure” described the same risk wearing different names. Sector labels alone don’t tell you what a portfolio is exposed to.

Worth keeping in proportion: Wall Street’s volatility gauge, the VIX, closed Tuesday at 15.84, its highest since August 4. That’s an elevated reading for a calm stretch, not a panicked one. This was a repricing, not a rout.

What this week is testing

3 events land in the next 5 days, and each one bears on the same question.

Nvidia’s quarterly results, with analysts expecting revenue near $92 billion. The number that matters more is the guidance, because it speaks to how much additional borrowing-financed AI capacity is coming, which loops straight back into bond supply.

A long aisle between two rows of server racks in a data center, lit evenly in white light.
Illustrative. The compute capacity behind results like these is what turns AI spending plans into actual bond issuance.

The core PCE price index, the Fed’s preferred inflation measure. A cooler reading eases the inflation half of the term-premium problem. A warmer one doesn’t.

Kevin Warsh’s first Jackson Hole speech as Fed Chair, on Friday. Traders want to know how he reads the yield move and whether he’ll address it directly. Most expect no change at the September meeting, and the odds of a hike sit near 1 in 3, at roughly 31.6%. Note what that number did not do: on August 17 we reported those same odds near 30%. A week of the sharpest bond selling since 2007 barely moved them, which says the market read the yield surge as a story about debt supply and term premium rather than about what the Fed is likely to do next month.

A wide mountain valley in late summer with dry golden grassland and a forested ridgeline under a clear sky.
Illustrative. Jackson Hole, Wyoming, where the Fed's annual economic symposium is held each August.

The thread back to last week’s commentary is worth naming. On August 17 we wrote that markets had read a softening consumer as pressure coming off the Fed. Last week the bond market answered: long-term rates rose anyway. A central bank holding still doesn’t hold the whole curve still.

Risk and Context

Rising long-term yields carry real costs. Mortgages, car loans, business credit and government interest expense all take their cue from the long end of the curve, not the fed funds rate. That’s the transmission from a bond auction to a household budget.

Bond prices and yields move in opposite directions, so a rising-yield stretch means existing bondholders see the market value of their holdings fall, even though a bond held to maturity still pays its stated interest and principal. Both facts are true at once, and the pairing confuses people constantly.

Single-week index moves say little about longer horizons, and the individual stock moves cited here illustrate the week’s mechanics rather than any view about those companies. One quarter of results from one chipmaker doesn’t determine a sector’s direction. Geopolitical situations like the US-Iran standoff can reverse quickly in either direction, and oil-linked inflation expectations move with them.

For the investor who reads the balance sheet, and the one who just wants to know what happened. Here at VEM, we tailor to both.

Sources

By Vann Equity Management

← Back to Blog