There’s a number that used to matter a great deal in portfolio construction, and it has been quietly ignored for most of the last 15 years. It’s the equity risk premium: the extra return investors expect to earn for owning stocks instead of simply buying government bonds.

Right now, that number has nearly disappeared.

The S&P 500’s trailing earnings yield sits at roughly 3.98% as of late June. The 10-year Treasury note is yielding around 4.37% to 4.40%. That puts bonds ahead of stocks on a realized basis by roughly 40 to 45 basis points, hovering near the narrowest gap since the dot-com era. Translation: safe government paper is paying more than equities, on a trailing basis, for the first time in over two decades.

And Treasuries, unlike stocks, don’t require you to be right about earnings growth, margin expansion, or AI monetization timelines.

This matters for a specific reason. For the better part of 2009 through 2021, stocks always won the yield comparison. Rates were so low that there was no real competition. The TINA trade, There Is No Alternative, drove a 15-year run that pushed S&P multiples into the 26x to 30x range without serious pushback from bond math. That math is now fundamentally different.

The counterargument is forward earnings. Using projected rather than trailing earnings, the picture shifts somewhat. The current FactSet consensus calls for S&P 500 earnings to grow roughly 24% in 2026 and about 17% in 2027, with a forward P/E sitting around 20x. If that holds, the forward earnings yield becomes more competitive. Goldman Sachs has raised its 2026 EPS forecast to $340, representing 24% annual growth, with AI infrastructure beneficiaries expected to account for roughly half of that earnings gain.

Here’s the thing, though. The companies driving that growth are largely spending from cash flow rather than borrowed money, which makes them less sensitive to rate increases than traditional rate-sensitive sectors. That’s a real distinction. But it applies to a fairly narrow slice of the index.

Slight tangent, but it matters: the rate picture has gotten more complicated than most people expected entering 2026. Goldman Sachs now expects the Fed to keep rates unchanged through all of 2026, with the first cuts pushed to June and December 2027. That is the third time this year Goldman has delayed its rate-cut timeline. The catalyst was a May jobs report that came in at 172,000 nonfarm payrolls, roughly double the consensus estimate. Meanwhile, markets are currently pricing in about a 63% probability of a rate hike at the September meeting.

So the bond market isn’t just holding yields high. It’s pricing in the possibility that yields go higher still.

What that means in practice: the investment-grade corporate bond market, which offers spreads above Treasuries, is now genuinely competitive with dividend-paying equities on a risk-adjusted basis for the first time since the early 2000s. Short and intermediate duration bonds are yielding enough to matter in a real portfolio context. That wasn’t true for most of the last decade.

The part people keep skipping is what this does to relative sector positioning inside equities. When bonds offer a real alternative, the sectors that benefit most are the ones with bond-like characteristics: predictable cash flows, strong balance sheets, pricing power. The sectors that lose are the ones pricing in multiple years of future earnings at premium multiples with no current dividend support.

Market action in 2026 has started to reflect this. The Russell 2000 is up roughly 19% year-to-date, compared with around 8% for the cap-weighted S&P 500. That broadening isn’t random. It tends to happen when capital starts flowing toward value and away from a concentrated bet on a handful of mega-caps that have already priced in the best-case outcome. The S&P 500’s top ten names now account for roughly 37% to 38% of the entire index by weight, a level of concentration that makes broad beta increasingly exposed to a small number of headlines.

None of this means the bull market is finished. The earnings growth underpinning this market is real in a way it wasn’t during the zero-rate era. First-quarter earnings grew 18% year over year, and the median S&P 500 company posted its strongest quarterly growth in roughly a decade, outside of the post-pandemic reopening period. That’s not nothing.

But the old portfolio math, hold concentrated tech, ignore bond competition, trust the AI premium, is operating in a very different rate environment than the one that made it work. If you haven’t stress-tested your equity holdings against a world where Treasuries pay 4.4% and stay there through mid-2027, now seems like a reasonable time to start.