Defensive Stocks Are Where A Five Percent Ten Year Does Its Damage

The question being asked everywhere this week is whether a record S&P 500 can survive a ten-year Treasury yield at a nineteen-year high. It is the wrong question. The index can survive it. What cannot survive it is the corner of the equity market that American investors have spent two decades treating as the safe one: the regulated, the leveraged and the high-yielding. Utilities, real estate and the dividend-income complex are not shelter from a 5 percent long bond. They are the asset most directly in its path.

So the right response to what the bond market did this week is not to cut equity exposure. It is to cut equity duration. The uncomfortable part of that argument, and the part reasonable people will dispute, is that the mega-cap growth cohort which textbook discounted cash flow analysis calls the longest-duration asset in the index is close to the last thing you should be selling here.

What the bond market actually did

On Wednesday 23 September the ten-year Treasury yield jumped 17 basis points to 5.12 percent, its highest since 2007, and it touched 5.18 percent before the week was out. The two-year note yield rose above 4.94 percent. Most importantly for what follows, the thirty-year bond yield reached 5.501 percent, a level last seen in June 2004.

That thirty-year number is the one that matters and the one nobody leads with. A twenty-two-year high in the long bond is not a statement about next month's Federal Reserve meeting. It is the discount rate against which a utility's rate base and a property portfolio's cap rate are valued.

The proximate causes were a manufacturing survey showing the greatest recorded increase in activity since early 2015, with input costs rising at their steepest rate in four years, and oil that will not come down. Brent settled at 104.32 dollars a barrel on Friday and West Texas Intermediate at 92.41 dollars.

Utilities already told you, and it was misread as a rotation

Here is the evidence that this is already happening rather than a forecast. By the time the ten-year had climbed to around 4.8 percent, the utilities sector's year-to-date gain had narrowed to roughly 2.3 percent while the S&P 500 was up more than 11 percent. On Wednesday's down session the Utilities Select Sector fund fell 1.9 percent, the worst of any sector, against an index that fell 0.8 percent.

That is not a rotation, and the arithmetic behind it is not complicated. The S&P 500 utilities sector carried a trailing twelve-month dividend yield of roughly 3.15 percent as of July. A two-year Treasury note pays 4.94 percent with no equity risk, no regulatory risk and no capital expenditure program to fund. For an income buyer the comparison is over. Either the share price falls until the dividend yield competes, or the dividend grows faster than regulators will allow.

Most commentary has explained the sector's 2026 underperformance away as the unwinding of a crowded artificial intelligence power-demand trade. Some of it was. But the problem above is simpler, and it applies whether or not the data centers need the electricity.

Real estate is the same trade, one step behind

This is where the argument becomes contestable, so let me state it plainly. Real estate has not repriced yet, and the consensus has drawn the wrong conclusion from that fact.

As of the end of August, the S&P 500 real estate sector had returned 14.66 percent year to date against 12.97 percent for the index, and the broader FTSE EPRA Nareit USA index 19.53 percent. REIT credit spreads stayed contained. The bull case reads this as proof that property has decoupled from rates.

I think that confuses a lag for a decoupling. Real estate absorbed the ten-year going from 3.9 percent in February to 4.7 percent in May for one reason: everybody still believed the next move in the policy rate was down. A bond proxy carries a higher discount rate perfectly well when that rate is expected to fall back to meet it. What it cannot carry is a discount rate going the other way on purpose.

The sixteenth of September removed the assumption

On 16 September the Federal Reserve raised the federal funds target range by 25 basis points to 3.75 to 4.00 percent. The vote was 12 to 0. It was the first increase since 2023. By the end of last week markets were putting roughly two-thirds odds on another 25 basis points in October, and officials' own year-end projections sat between 4.1 and 4.4 percent.

That is the regime change, and it is why utilities' underperformance should not stop at utilities. Every bond proxy in the index was underwritten on an assumption about the direction of policy that stopped being true nine days ago. Utilities are what the removal of that assumption looks like once it has been priced. Real estate is what it looks like before.

Why the mega caps are not the long-duration asset here

The standard objection is that high-multiple growth is the most rate-sensitive thing in the index, because more of its value sits in distant cash flows. As algebra that is true. In this particular cycle it is beside the point, for two reasons.

The first is funding. A regulated utility or a property company must go to the capital markets to fund the asset base that generates its earnings, and must now do it at 5.5 percent thirty-year money rather than the 3 percent money those assets were modeled on. The largest technology companies fund capital expenditure out of operating cash flow and sit on net cash. A higher risk-free rate raises the hurdle they must clear. It does not raise their cost of doing business in the same mechanical way.

The second is the numerator. What is pushing yields up is nominal: input costs, fuel, transport, a manufacturing sector booming in price terms. Companies whose revenues reprice annually get some of that back in earnings. Companies whose revenues are fixed by a regulator's allowed return, by a ten-year lease signed in 2021, or by a coupon, get none of it. Inflation that lifts the discount rate and the cash flows together is survivable. Inflation that lifts only the discount rate is not.

Friday showed this. The Dow rose 0.93 percent to 51,828.62, snapping a three-week losing streak, the S&P 500 gained 0.51 percent to 7,743.41 and the Nasdaq Composite 0.5 percent to 27,068.72. On the week the S&P added 0.63 percent and the Nasdaq 1.21 percent, having set an all-time closing high along the way. Mega-cap technology carried the index through the worst bond week in nearly two decades.

The consumer problem lands in the same place

A second reason for wariness is the household data. The University of Michigan's September sentiment index finished at 48.1, near historic lows and below the 51.0 expected. Year-ahead inflation expectations rose to 4.6 percent from 4.0 percent in August, the highest since June. Views of both current and year-ahead personal finances deteriorated by around 10 percent in the month.

Set that against the price data. August CPI, released on 11 September, put headline inflation at 3.4 percent year on year, unchanged from July, with a monthly rise of 0.4 percent driven substantially by gasoline. Core inflation was 2.4 percent, the lowest since March 2021.

Read together, those give the shape of the problem. Underlying inflation has essentially arrived. What is left in the headline is an energy price set in the Strait of Hormuz, which no amount of tightening in Washington can reach. A central bank hiking against that will not be validated quickly by core, which means this cycle is likelier to end through damage to demand than through visible disinflation.

Damage to demand lands on domestic, leveraged, consumer-facing balance sheets. Energy is the best sector in the index this year at 47.7 percent; consumer discretionary is the worst at minus 5.0 percent, and it was the second-weakest sector on Wednesday at minus 1.5 percent. That is a market that has already worked out where the pain goes.

What would change my mind

This is a falsifiable view, so here is what would break it.

• Core inflation back above 3 percent. If the energy shock feeds into services and wages, this stops being a duration story and becomes a broad valuation story. At a forward price to earnings ratio of 22.1 against a ten-year average of 18.8, the index earns roughly 4.5 percent forward against a 5.18 percent Treasury. In that world the mega-cap multiple breaks too and nothing in equities is shelter.
• Brent back toward 70 dollars with the Fed signaling a pause. That restores the falling-rate assumption the income sectors were underwritten on. The argument rests on the direction of policy, not its level.
• Mega-cap free cash flow turning negative. If the capital expenditure cycle grows large enough that the biggest technology companies must fund it externally, the self-funding premise disappears and they become as rate-sensitive as everybody else.
• Regulators moving faster than expected. Allowed returns are not fixed forever. If utility regulators reset them upward in line with a 5.5 percent long bond quicker than the historical pattern suggests, the sector's repricing is nearer its end than its beginning.

Absent those, the position holds. A nineteen-year high in the ten-year and a twenty-two-year high in the thirty-year are not arguments for holding less equity. They are arguments for holding equity whose cash flows move with the thing causing the yields to rise, and for being honest that the word defensive no longer describes what a great many American portfolios own.