Will bonds finally rally?

Credit where it’s due: the bond bulls are reacting to something real. A 10-year yield near 5% is the highest entry in fifteen years, the “everyone’s short” headlines are backed by genuinely record futures positioning, and it’s true that China and Japan aren’t the buyers they used to be. Those are the right things to notice. The question is whether they add up to “bonds are about to rally.”

So I took each leg one step past the slogan — tested the mean-reversion claim on sixty-four years of yields, looked at what the “record short” actually is, pulled the Treasury’s own data on who owns the debt, and then asked the only question that matters for a portfolio: given all that, what do you actually hold? The short version: the bull case survives, but for a different reason than the one being sold — and in today’s regime, plain cash expresses it better than long bonds.

01Do yields actually mean-revert?

This is the heart of the “they’re too high” argument, and it borrows a mechanism from stocks. Equities mean-revert on valuation: a cheap decade tends to be followed by a rich one. The instinct is to assume bonds do the same — a high yield is “cheap” and should snap back. But a yield has no fixed level to snap back to, and the data says so plainly. Run an augmented Dickey-Fuller test on the 10-year yield and it fails to reject a unit root in every sub-period — 1962–81, 1981–2020, and 2020–today. In plain English: statistically it’s a near-random walk that trends for decades, not a spring. The full-sample “half-life” of reversion is about nine years, which is another way of saying there’s nothing tradeable there.

What bonds do have is the cleanest relationship in all of macro: your forward return is basically the yield you start with. Build a rolling 10-year Treasury and the correlation between its starting yield and the next decade’s annualized return is +0.94 — the starting yield alone explains 88% of what you earn. So the grain of truth in “bad decade → good decade” is real, but the mechanism is carry, not reversion: the bad decade happened because yields rose, and the higher yield is the future return. At ~5% today you have a decent starting point — but that’s an argument about carry, not about yields being “due” to fall.

02“Everyone’s short, so a squeeze is coming”

The positioning really is at a record — but it’s worth knowing what the position is before betting on a squeeze. The giant “short Treasuries” figure is overwhelmingly the basis trade: hedge funds buy the cash bond and sell the matching futures contract, pocketing the tiny price gap between them plus the carry, financed in the repo market and levered 50–100× to make the pennies add up. The Fed estimates Cayman-based funds alone hold around $2 trillion of Treasuries this way.

The key point: that book is hedged. The short-futures leg is offset by the long-cash leg, so it’s roughly neutral to the level of rates — it doesn’t profit if yields rise and won’t panic-cover into a rally if they fall. It isn’t a pile of directional bears waiting to be squeezed. And the ETF version tells the same story: short interest in the big long-bond ETF peaked around 27% of float late in 2025 and has since drifted down to ~19% — the crowded short quietly came off, and there was no squeeze. The bond simply kept drifting while the shorts covered.

03Are foreigners fleeing — and if so, who’s buying?

This is the best of the three instincts, and it’s half right. China is selling: its holdings are down about a third over the decade to an 18-year low, a deliberate reserve-management retreat (and it’s rotated toward gold). Japan trimmed recently too. But step back and the “foreigners are dumping” story needs two corrections. First, in dollars there’s no dumping — total foreign holdings are near an all-time record around $9.3 trillion. What fell is their share of the market, from ~34% in 2013 to ~24% now — and that share actually ticked back up the last two years. The U.S. issued debt faster than foreigners bought it; it didn’t lose them.

So who is absorbing all the new issuance, with auctions still clearing? The buyer base shifted from price-insensitive foreign central banks to price-sensitive domestic money. The Fed went from buyer to seller under quantitative tightening. In its place: money-market funds (swollen past $7 trillion, voracious buyers of Treasury bills), U.S. households buying directly and through funds, and banks (back up to $1.8 trillion from $700 billion pre-pandemic). Domestic and foreign private investors are now about 60% of the market, up from 37% in 2014. Treasury leaned issuance toward short bills, which the money funds swallow easily — which is exactly why auctions don’t fail.

04So what do you actually hold?

Here’s where the rabbit hole pays off. If yields don’t mean-revert and the short isn’t squeezable, the real reason to own bonds is the boring one: you’re paid ~5% to wait, and they’re insurance against a recession — the one scenario that would hurt a stock-heavy portfolio most. But that insurance is regime-dependent, and right now the regime is working against it. Stock/bond correlation, reliably negative from 2008–2020, flipped positive in 2022 and sits around +0.3 today. When the shock is inflation rather than growth, long bonds fall with stocks — or worse.

I pressure-tested this on a real, stock-heavy book (a ~80% equity sleeve plus a 20% defensive sleeve, rebalanced monthly). Over 2011–2026, swapping that 20% between cash, intermediate bonds, and long bonds barely moves the headline numbers — but where it matters, the tail, cash wins the current regime: it gave the best risk-adjusted return and the smallest drawdown, won 2022 and 2025–26, and gave up only about a point to duration in COVID. And the carry you’d be reaching for is a mirage after tax: T-bills and intermediate Treasuries yield almost the same (~4.3% vs ~4.4%), both taxed the same way, so there’s no income reason to take the duration risk. Long bonds only win if yields fall — which is the recession bet, not today’s inflation risk.

05The honest bottom line

Bonds can rally — but not for the reasons on the marketing sheet. Not because yields mean-revert (they don’t; they trend for decades). Not because a crowded short gets squeezed (it’s a hedged arbitrage, not a bet). And not because foreigners are dumping (in dollars they’re near a record; the pressure is supply, which if anything pushes yields up). The one path that sends yields meaningfully lower is a growth shock that tips the economy into recession — and ironically the oil spike pushing yields up today is exactly the kind of thing that could cause it a year out. That’s the real bull case: not a coiled spring, but cheap insurance against the thing you’re most exposed to — bought, for now, more cleanly in cash than in long bonds.

None of this makes the bond bulls wrong to be interested — a 5% yield is the most interesting bonds have been in years, and the instinct to look at positioning and foreign flows is the right instinct. Chasing each leg one step further just changes the why: the case is carry and recession insurance, not mean reversion or a squeeze, and the cleanest way to hold it in an inflation-prone, flat-curve world is the most boring one — cash, with a measured dash of duration for the tail.