The 10-Year Treasury Is Back Above 5%. Inflation Decided Who Won the Last Seven Times.

Ruth has a CD maturity notice on her kitchen table and a question she has been asking everyone for a week. She is 63, and the $50,000 in that CD is money she expects to spend in her early 70s. A 10-year Treasury note pays 5.11%. Her online savings account pays 4.20%. The headlines say yields are the highest since 2007, and she cannot tell whether that means "lock it in now" or "wait, it is going higher." (Ruth is a composite, built to show a typical decision; the numbers around her are real.)

The coverage she has been reading splits cleanly. Market pieces describe a bond rout. Income pieces say 5% is a gift. Almost none of them answer the question she is actually asking: the last few times the 10-year got here, what happened to the people who locked in, compared with the people who waited?

That has a findable answer. Since 1962 there have been seven episodes that look like this one, where the 10-year yield had just climbed at least three-quarters of a point in a year and landed somewhere between 4% and 6.5%. In four of them, locking in won. In three, waiting won. What separated the two groups was what happened to inflation afterward. The Fed and the level of the yield mattered less.

Where yields stand on September 23, 2026

The 10-year Treasury closed at 5.11% on September 23, up 15 basis points in a single session (a basis point is one-hundredth of a percentage point) and its highest close since July 13, 2007, per the U.S. Treasury's daily yield curve. A year earlier it was 4.12%. The 30-year closed at 5.40%, its highest since July 2004. The 2-year sat at 4.85%, and the 3-month bill at 4.19%.

The backdrop is one the market has not seen in a while. The Bureau of Labor Statistics reported on September 11 that consumer prices rose 3.4% in the year through August, or 2.4% excluding food and energy. Five days later, the Federal Reserve raised its target range by a quarter point to 3.75% to 4.00% on a 12-0 vote, with a statement that said plainly: "Inflation remains elevated."

Cash is still paying, too. The best high-yield savings account SwitchWize tracks paid 4.20% on September 24, and the best 5-year CD paid 4.50%. So Ruth is choosing between three reasonable options, and none of them is bad. The question is which one history favors from here.

How rare a return to 5% is

The 10-year has spent a lot of its history above 5%. It averaged about 5.8% at month-end from 1962 through this September. What is rare is coming back to 5% after years below it. Using closing prices from FRED's DGS10 series, that has happened three times: February 28, 1966; April 13, 2006; and September 15, 2026. The October 2023 run got close, topping out at a 4.98% close on October 19, then reversing.

10-year Treasury yield, quarterly averages, with the latest point set to the September 23, 2026 close. Source: FRED (DGS10), U.S. Treasury.

Those two earlier crossings are almost perfect opposites, which is what makes them useful.

In 1966, the 5% crossing was the start of something. The Federal Reserve History essay on the Great Inflation dates that era from 1965, when federal spending on the Vietnam War and new domestic programs ran up against a Fed that let money growth run. Inflation, which had been a little over 1% a year in 1964, was 2.6% by February 1966 and 4.7% by February 1969. The 10-year followed it: 5.02% at the end of February 1966, 6.26% three years later, and 8.22% by May 1970.

In 2006, the crossing was close to the end. The 10-year touched a 5.26% close in June 2007, then the housing bust and the 2008 financial crisis pulled it down to 3.16% by April 2009. The 3-month bill went from 4.65% to 0.14% over the same three years.

What $50,000 earned in seven comparable episodes

Two crossings are not enough to learn from, so SwitchWize widened the test. The rule: find every month since 1962 where the 10-year had risen at least 0.75 percentage points over the prior year and closed between 4% and 6.5%. That filter produces seven distinct episodes. For each one, the test compares two savers with $50,000 on the first qualifying month:

• The locker buys a 10-year Treasury note at par, collects its interest, parks that interest in T-bills, and after three years is valued at what the note would sell for.
• The roller keeps the whole $50,000 in 3-month T-bills, rolling them over as they mature at whatever the going rate is.

The locker won four times and lost three. The wins were bigger than the losses. The best case, April 2006, left the locker $8,957 ahead ($13,795 against $4,838). The worst, February 1966, left the locker $3,218 behind ($4,643 against $7,861).

Inflation is the trailing 12-month change in CPI-U (FRED, CPIAUCSL). October 2025 CPI was never published because of that fall's federal government shutdown, so the Oct 2022 row's end reading uses September 2025.

Why inflation decided the outcome

The two savers are exposed to different things, and the last column of the table shows how.

The roller's income resets every three months. When inflation keeps the Fed raising or holding short-term rates, the roller's T-bills pay more almost immediately. In 1966, bills went from 4.64% to 6.19% in three years. The roller was carried upward by the same inflation that was eroding everyone's money.

The locker's income is fixed for ten years, and the locker loses in one of two ways. The first is price: rising inflation pushes new 10-year yields higher, which lowers what the old note would sell for. From February 1966 to February 1969, inflation went from 2.6% to 4.7% and the 10-year went from 5.02% to 6.26%, and the markdown on the note outweighed its higher income. The second is income: if inflation stays high enough that the Fed keeps short rates above the note's yield, the roller simply earns more each month. That is what happened after October 2022. Inflation was falling, from 7.8% to about 3%, but it stayed above the Fed's 2% target, and T-bills paid more than the 4.10% note for 33 of the next 36 months.

Run the same logic in reverse and the wins make sense. In all four, inflation was lower three years later than at the start. In 1999 and 2006 a recession arrived inside the window, the Fed cut short-term rates hard, and the roller's income collapsed while the locker's stayed put. In April 2006, bills paid more than the note in none of the next 36 months.

That is why "is 5% the top?" is the wrong question. Yields can keep rising and the locker can still win if they rise slowly enough; they can fall and the roller can still do fine if the Fed holds short rates up for long enough. The question that sorted these seven episodes is whether inflation over the next few years eases, or rises and stays high enough to keep the Fed pushing short rates above the note.

What the market is pricing in now

The market's own guess at future inflation is visible in the gap between an ordinary Treasury and an inflation-protected one (a TIPS, whose principal rises with consumer prices). On September 23, the 10-year TIPS yielded 2.76%, per the Treasury's real yield curve. Subtract that from the 5.11% nominal yield and the market is pricing about 2.35% a year of inflation over the next decade.

So a locker at 5.11% comes out ahead of the inflation-protected alternative if inflation averages less than about 2.35% over the life of the note, and behind it if inflation averages more. Headline inflation is running at 3.4% today. Core inflation, which strips out food and energy, is 2.4%, right on the market's number.

The market is betting today's energy-driven inflation spike fades. The 1966 lockers made the opposite mistake. They bought in a year when inflation was also 2.6% and looked contained, and it did not stay contained.

There is one real difference in Ruth's favor. The 10-year TIPS paying 2.76% above inflation did not exist in 1966 (the Treasury first sold TIPS in 1997). Today a saver worried about a repeat of the late 1960s can buy that protection directly instead of guessing. The Real Yield Index tracks what savings and Treasury products pay after inflation.

Why the record is thinner than it looks

Seven episodes is a thin record, and it is honest to say so. Two of the three losses come from the same inflationary era, the late 1960s, which means the history is really four wins, one inflationary regime that produced back-to-back losses, and one recent loss. The wins are not independent either: two of them were rescued by recessions, and nobody should buy a bond hoping for a recession.

The broader record is less flattering to lockers than the seven-episode test. Across every month since 1962 when the 10-year was up at least 0.75 points on the year, at any starting level, a buy-and-hold note beat T-bills over the following 12 months about half the time and over the following 36 months about two-thirds of the time. Those months overlap heavily, so they are not 166 separate experiments. They do show that locking in after a sharp rise has been a modest edge over three years, not a sure thing, and close to a coin flip over one.

And the starting yield matters more than any episode rule. The worst case in the full record came in March 2021, after the 10-year had jumped from 0.6% to 1.74%. A $50,000 note bought then was worth $4,626 less three years later, while T-bills earned $4,234. A 1.74% yield had no income to absorb the 2022 selloff. A 5.11% yield does.

How much room a 5.11% note has

That income cushion can be measured. A 10-year note bought at 5.11% pays about $2,555 a year on $50,000. If yields rise over the next twelve months, the note's price falls, and the question is how far they can climb before the price loss eats the year's interest.

SwitchWize calculation: standard semiannual bond pricing, par note held one year, interest included. A roller earning today's 4.19% bill rate would make about $2,095.

A rise of about three-quarters of a point in a year is where the locker breaks even. That is not rare: in about 22% of the rolling 12-month windows since 1962, the 10-year rose by more than that. But it is also not the likely case, and the locker who does not sell never realizes that paper loss at all.

That last point matters more than any chart. A Treasury note held to maturity pays exactly what it promised: 5.11% a year, then $50,000 back. The price swings in the table only hit someone who sells before then. The locker's real risk is not a loss of dollars. It is that inflation erodes what those dollars buy, or that a better rate comes along while their money is committed. The cash duration guide walks through how to size that commitment to the date you actually need the money.

How the 30-year compares

Stretching to the 30-year for its extra 0.29 points of yield buys much less cushion. Using the same math, a 30-year bond bought at 5.40% shows a loss if its yield rises about 0.4 points in a year, roughly half the 10-year's room, and a one-point rise leaves it down about 7.7% even after a year of interest. The longer the bond, the more one year's price move outweighs one year's interest. SwitchWize's piece on the 30-year Treasury's hidden catch covers that trade in detail.

What the history says to do with a 5% yield

The seven episodes point to a few rules that hold whichever way inflation breaks:

1. Match the maturity to the date you need the money. Money needed in 2033 can sit in a note maturing in 2033 and ignore every price swing between now and then. Money you might need next spring belongs in bills or savings, where a rate reset is a feature.
2. Treat a long fixed rate as an inflation forecast. At 5.11%, the break-even is inflation averaging about 2.35% for ten years. If that sounds too low to you, a TIPS makes the bet for you at 2.76% above inflation.
3. Split the decision instead of timing it. None of the seven lockers knew which episode they were in. A saver who put half into a note and half into bills would have landed between the two outcomes every time, never at the extremes.
4. Remember that Treasury interest skips state income tax. Bills and notes are exempt from state and local income tax; savings account and CD interest is not. In a high-tax state, that is enough to put a 4.19% bill ahead of a 4.20% savings account after tax. The CD, bond and Treasury comparison runs the after-tax math.

For current bank rates to compare against, see the best CD rates and high-yield savings rates SwitchWize tracks daily.

Ruth read the table twice and decided she was not going to bet on which decade this turns out to be. She moved $30,000, the part she expects to spend at 70 and 71, into a 7-year Treasury note paying 5.05% that matures the year she needs it, so the price swings in between will never touch her. She left the other $20,000 in 3-month bills, rolling them, so that if inflation does run like 1966, part of her money resets with it.

Quick answers

Is 5% on the 10-year high by historical standards? It is the highest close since July 2007, but the 10-year has averaged about 5.8% at month-end since 1962. What is rare is the return to 5% after years below it, which has happened only in 1966, 2006 and 2026.

Did locking in at 5% pay off in the past? In four of seven comparable episodes since 1966, a 10-year note held for three years beat rolling T-bills. It lost when inflation rose or stayed high enough to keep T-bills paying more (1966, 1967, 2022) and won when inflation eased.

What is the break-even for a note bought at 5.11%? About a 0.75-point rise in yields over the next year before the price drop cancels a year of interest. Held to maturity, the note pays 5.11% a year regardless.